diff --git a/apps/worker/video_processing/render_adapter.py b/apps/worker/video_processing/render_adapter.py
index bf20d9cc1..04a90dabe 100644
--- a/apps/worker/video_processing/render_adapter.py
+++ b/apps/worker/video_processing/render_adapter.py
@@ -485,6 +485,7 @@ class RenderAdapter:
         rendered_clip_ids: list[str] | None = None,
         failed_clip_ids: list[str] | None = None,
         voiceover_audio_path: str | None = None,
+        is_preview: bool = False,
     ) -> RenderAdapterResult:
         """执行统一渲染核心流程（BGM + ASR + 渲染 + 缩略图 + 上传）。
 
@@ -503,8 +504,8 @@ class RenderAdapter:
 
         self._report_progress(progress_cb, 40.0, "执行视频渲染")
 
-        # 2. 初始化 ASR
-        asr_service = self._get_asr_service()
+        # 2. 初始化 ASR（预览模式跳过，节省启动开销）
+        asr_service = None if is_preview else self._get_asr_service()
 
         # 3. 读取输出分辨率
         plan_config = plan.config or {}
@@ -529,25 +530,27 @@ class RenderAdapter:
             bgm_path=bgm_path,
             asr_service=asr_service,
             voiceover_audio_path=voiceover_audio_path,
+            is_preview=is_preview,
         )
         result = render_svc.render()
 
-        # 4.5 渲染后校验输出完整性
-
-        validation = validate_video_output(result.output_path)
-        if not validation.valid:
-            logger.error(
-                "[render-adapter] 渲染输出校验失败: plan_id=%s job_id=%s error=%s",
-                plan_id,
-                job_id,
-                validation.error_message,
-            )
-            return RenderAdapterResult(
-                success=False,
-                error_message=f"渲染输出校验失败: {validation.error_message}",
-                error_detail=validation.error_message,
-            )
-
+        # 4.5 渲染后校验输出完整性（预览模式跳过，节省耗时）
+        if is_preview:
+            logger.info("[render-adapter] 预览模式：跳过输出校验")
+        else:
+            validation = validate_video_output(result.output_path)
+            if not validation.valid:
+                logger.error(
+                    "[render-adapter] 渲染输出校验失败: plan_id=%s job_id=%s error=%s",
+                    plan_id,
+                    job_id,
+                    validation.error_message,
+                )
+                return RenderAdapterResult(
+                    success=False,
+                    error_message=f"渲染输出校验失败: {validation.error_message}",
+                    error_detail=validation.error_message,
+                )
         self._report_progress(progress_cb, 80.0, "上传渲染结果")
 
         # 5. 上传结果
@@ -556,19 +559,20 @@ class RenderAdapter:
 
         self._report_progress(progress_cb, 90.0, "生成封面缩略图")
 
-        # 6. 生成缩略图
+        # 6. 生成缩略图（预览模式跳过，节省耗时）
         thumbnail_url = ""
-        try:
-            from video_processing.thumbnail_generator import generate_and_upload_thumbnail
+        if not is_preview:
+            try:
+                from video_processing.thumbnail_generator import generate_and_upload_thumbnail
 
-            thumb_storage_key = f"rendered/{plan_id}/thumbnail.jpg"
-            thumbnail_url = generate_and_upload_thumbnail(str(result.output_path), thumb_storage_key)
-        except Exception as thumb_err:
-            logger.warning(
-                "[render-adapter] 缩略图生成失败（不影响主流程）: plan_id=%s error=%s",
-                plan_id,
-                thumb_err,
-            )
+                thumb_storage_key = f"rendered/{plan_id}/thumbnail.jpg"
+                thumbnail_url = generate_and_upload_thumbnail(str(result.output_path), thumb_storage_key)
+            except Exception as thumb_err:
+                logger.warning(
+                    "[render-adapter] 缩略图生成失败（不影响主流程）: plan_id=%s error=%s",
+                    plan_id,
+                    thumb_err,
+                )
 
         self._report_progress(progress_cb, 100.0, "渲染完成")
 
@@ -614,6 +618,7 @@ class RenderAdapter:
         work_dir: Path | None = None,
         progress_cb: ProgressCallback | None = None,
         voiceover_audio_path: str | None = None,
+        is_preview: bool = False,
     ) -> RenderAdapterResult:
         """使用内存中的 plan/clips/asset_path_map 直接渲染。
 
@@ -672,6 +677,7 @@ class RenderAdapter:
                 job_id=job_id,
                 progress_cb=progress_cb,
                 voiceover_audio_path=voiceover_audio_path,
+                is_preview=is_preview,
             )
 
         except subprocess.CalledProcessError as exc:
diff --git a/apps/worker/video_processing/unified_render_service.py b/apps/worker/video_processing/unified_render_service.py
index c8f6229c8..57d8e5b2e 100755
--- a/apps/worker/video_processing/unified_render_service.py
+++ b/apps/worker/video_processing/unified_render_service.py
@@ -151,6 +151,7 @@ class UnifiedRenderService:
         asr_service: Any = None,  # ASRService 实例，用于自动生成字幕
         bgm_path: str | None = None,  # BGM 本地文件路径
         voiceover_audio_path: str | None = None,  # 配音素材库音频本地路径
+        is_preview: bool = False,  # 预览模式：ultrafast 编码 + 跳过非必要步骤
     ):
         self.plan = plan
         self.clips = clips
@@ -163,6 +164,7 @@ class UnifiedRenderService:
         self.asr_service = asr_service
         self.bgm_path = bgm_path
         self.voiceover_audio_path = voiceover_audio_path
+        self.is_preview = is_preview
         self._transition_engine = TransitionEngine(default_duration=transition_duration)
         self._speed_engine = SpeedEngine()
         self._asr_timeline_cache: Any = None  # ASR 字幕结果缓存，避免重复调用
@@ -1228,9 +1230,9 @@ class UnifiedRenderService:
             "-c:v",
             "libx264",
             "-crf",
-            "23",
+            "28" if self.is_preview else "23",
             "-preset",
-            "medium",
+            "ultrafast" if self.is_preview else "medium",
             "-pix_fmt",
             "yuv420p",
             "-movflags",
@@ -1742,9 +1744,9 @@ class UnifiedRenderService:
             "-c:v",
             "libx264",
             "-crf",
-            "23",
+            "28" if self.is_preview else "23",
             "-preset",
-            "medium",
+            "ultrafast" if self.is_preview else "medium",
             "-pix_fmt",
             "yuv420p",
             "-movflags",
@@ -1753,10 +1755,11 @@ class UnifiedRenderService:
         ]
 
         logger.info(
-            "执行渲染: plan_id=%s inputs=%d output=%s",
+            "执行渲染: plan_id=%s inputs=%d output=%s preview=%s",
             self.plan.id,
             input_args.count("-i"),
             output_path,
+            self.is_preview,
         )
         try:
             run_ffmpeg(command)
diff --git a/apps/worker/worker_app/tasks/generation.py b/apps/worker/worker_app/tasks/generation.py
index d2c9ffb83..04ca03599 100644
--- a/apps/worker/worker_app/tasks/generation.py
+++ b/apps/worker/worker_app/tasks/generation.py
@@ -743,7 +743,8 @@ def _download_library_assets(
                     len(asset_ids),
                 )
 
-            downloaded: list[Path] = []
+            # 构建待下载列表 (index, asset, storage_key, local_file)
+            download_jobs: list[tuple[int, Any, str, Path]] = []
             failed_assets: list[str] = []
             for i, asset in enumerate(assets):
                 storage_key = asset.file_url if asset.file_url else None
@@ -772,52 +773,81 @@ def _download_library_assets(
 
                 ext = Path(storage_key).suffix or ".mp4"
                 local_file = temp_path / f"asset_{i:03d}_{asset.id}{ext}"
-                asset_start = time.monotonic()
-                download_ok = download_asset(storage_key, local_file)
-                asset_elapsed = time.monotonic() - asset_start
+                download_jobs.append((i, asset, storage_key, local_file))
 
-                if download_ok:
-                    file_size = local_file.stat().st_size if local_file.exists() else 0
-                    downloaded.append(local_file)
-                    logger.info(
-                        "[task_id=%s] Downloaded asset: %s -> %s (size=%d, time=%.1fs)",
-                        task_id,
-                        asset.name,
-                        local_file,
-                        file_size,
-                        asset_elapsed,
-                    )
-                    if gen_task:
-                        gen_task.append_log(
-                            "下载素材",
-                            f"下载成功: {asset.name}",
-                            asset_id=asset.id,
-                            asset_name=asset.name,
-                            success=True,
-                            file_size=file_size,
-                            duration=round(asset_elapsed, 2),
+            # 并行下载素材（线程池，IO 密集型）
+            downloaded: list[Path] = []
+            if download_jobs:
+                from concurrent.futures import ThreadPoolExecutor, as_completed
+
+                max_workers = min(len(download_jobs), 6)
+                logger.info(
+                    "[task_id=%s] 并行下载素材: count=%d, workers=%d",
+                    task_id,
+                    len(download_jobs),
+                    max_workers,
+                )
+
+                def _download_one(item: tuple) -> tuple[int, Any, Path, bool, float]:
+                    idx, asset, skey, lfile = item
+                    t0 = time.monotonic()
+                    ok = download_asset(skey, lfile)
+                    elapsed = time.monotonic() - t0
+                    return idx, asset, lfile, ok, elapsed
+
+                with ThreadPoolExecutor(max_workers=max_workers) as executor:
+                    futures = {executor.submit(_download_one, job): job for job in download_jobs}
+                    # 按原始顺序收集结果，保证 downloaded 列表顺序稳定
+                    results_map: dict[int, tuple[Path, bool, float, Any]] = {}
+                    for future in as_completed(futures):
+                        idx, asset, lfile, ok, elapsed = future.result()
+                        results_map[idx] = (lfile, ok, elapsed, asset)
+
+                # 按原始顺序处理结果
+                for idx in sorted(results_map.keys()):
+                    lfile, ok, elapsed, asset = results_map[idx]
+                    if ok:
+                        file_size = lfile.stat().st_size if lfile.exists() else 0
+                        downloaded.append(lfile)
+                        logger.info(
+                            "[task_id=%s] Downloaded asset: %s -> %s (size=%d, time=%.1fs)",
+                            task_id,
+                            asset.name,
+                            lfile,
+                            file_size,
+                            elapsed,
                         )
-                else:
-                    failed_assets.append(f"{asset.name}({asset.id})")
-                    logger.warning(
-                        "[task_id=%s] Failed to download asset: %s (id=%s)",
-                        task_id,
-                        asset.name,
-                        asset.id,
-                    )
-                    if gen_task:
-                        gen_task.append_log(
-                            "下载素材",
-                            f"下载失败: {asset.name}",
-                            level="WARN",
-                            asset_id=asset.id,
-                            asset_name=asset.name,
-                            success=False,
-                            file_size=0,
-                            duration=round(asset_elapsed, 2),
+                        if gen_task:
+                            gen_task.append_log(
+                                "下载素材",
+                                f"下载成功: {asset.name}",
+                                asset_id=asset.id,
+                                asset_name=asset.name,
+                                success=True,
+                                file_size=file_size,
+                                duration=round(elapsed, 2),
+                            )
+                    else:
+                        failed_assets.append(f"{asset.name}({asset.id})")
+                        logger.warning(
+                            "[task_id=%s] Failed to download asset: %s (id=%s)",
+                            task_id,
+                            asset.name,
+                            asset.id,
                         )
-                    if strict:
-                        raise RuntimeError(f"素材下载失败: asset_id={asset.id}, name={asset.name}")
+                        if gen_task:
+                            gen_task.append_log(
+                                "下载素材",
+                                f"下载失败: {asset.name}",
+                                level="WARN",
+                                asset_id=asset.id,
+                                asset_name=asset.name,
+                                success=False,
+                                file_size=0,
+                                duration=round(elapsed, 2),
+                            )
+                        if strict:
+                            raise RuntimeError(f"素材下载失败: asset_id={asset.id}, name={asset.name}")
 
             # 指定了 asset_ids 但全部下载失败 → 无论 strict 与否都报错
             if asset_ids and not downloaded:
@@ -1171,6 +1201,7 @@ def _render_video(
             job_id=task_id,
             work_dir=temp_path,
             voiceover_audio_path=voice_path,
+            is_preview=is_preview,
         )
     finally:
         db.close()
diff --git a/tests/unit/test_1280_preview_speedup.py b/tests/unit/test_1280_preview_speedup.py
new file mode 100644
index 000000000..13cb516cd
--- /dev/null
+++ b/tests/unit/test_1280_preview_speedup.py
@@ -0,0 +1,289 @@
+"""#1280 预览视频生成加速 — 单元测试。
+
+验证点：
+1. UnifiedRenderService.is_preview 参数正确传递
+2. 预览模式使用 ultrafast preset + crf 28
+3. RenderAdapter.render_from_memory 正确传递 is_preview
+4. 预览模式跳过 ASR 初始化
+5. 预览模式跳过输出校验和缩略图
+6. generation.py 并行下载逻辑
+"""
+
+from __future__ import annotations
+
+import os
+import tempfile
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+# ── 1. UnifiedRenderService is_preview 参数 ──
+
+
+class TestUnifiedRenderServicePreviewFlag:
+    """is_preview 参数正确传递和存储。"""
+
+    def test_default_is_preview_false(self):
+        from video_processing.unified_render_service import UnifiedRenderService
+
+        svc = UnifiedRenderService(
+            plan=MagicMock(id="test"),
+            clips=[],
+            asset_path_map={},
+            work_dir=Path(tempfile.mkdtemp()),
+        )
+        assert svc.is_preview is False
+
+    def test_is_preview_true(self):
+        from video_processing.unified_render_service import UnifiedRenderService
+
+        svc = UnifiedRenderService(
+            plan=MagicMock(id="test"),
+            clips=[],
+            asset_path_map={},
+            work_dir=Path(tempfile.mkdtemp()),
+            is_preview=True,
+        )
+        assert svc.is_preview is True
+
+    def test_is_preview_false_explicit(self):
+        from video_processing.unified_render_service import UnifiedRenderService
+
+        svc = UnifiedRenderService(
+            plan=MagicMock(id="test"),
+            clips=[],
+            asset_path_map={},
+            work_dir=Path(tempfile.mkdtemp()),
+            is_preview=False,
+        )
+        assert svc.is_preview is False
+
+
+# ── 2. 预览模式 FFmpeg 参数 ──
+
+
+class TestPreviewFFmpegPreset:
+    """预览模式使用 ultrafast preset + crf 28。"""
+
+    def _make_clip(self):
+        from video_processing.unified_render_service import ResolvedClip
+
+        return ResolvedClip(
+            clip_id="c1",
+            asset_id="a1",
+            local_path=Path("/tmp/fake.mp4"),
+            clip_type="main",
+            order=0,
+            start_time=0,
+            duration=10.0,
+            playback_speed=1.0,
+            transition_effect="cut",
+            transition_duration=0.0,
+            config={},
+        )
+
+    @patch("video_processing.unified_render_service.run_ffmpeg")
+    def test_execute_ffmpeg_preview_uses_ultrafast(self, mock_run):
+        from video_processing.unified_render_service import (
+            RenderLayer,
+            UnifiedRenderService,
+        )
+
+        plan = MagicMock()
+        plan.id = "test_plan"
+        plan.config = {"export": {"resolution": "854x480"}}
+
+        clip = self._make_clip()
+
+        svc = UnifiedRenderService(
+            plan=plan,
+            clips=[clip],
+            asset_path_map={"a1": Path("/tmp/fake.mp4")},
+            work_dir=Path(tempfile.mkdtemp()),
+            output_width=854,
+            output_height=480,
+            is_preview=True,
+        )
+
+        layers = [RenderLayer(role="main", clips=[clip])]
+        filter_complex, input_args = svc._build_filter_complex(layers)
+        output_path = Path(tempfile.mkdtemp()) / "out.mp4"
+        svc._execute_ffmpeg(filter_complex, input_args, output_path)
+
+        mock_run.assert_called_once()
+        cmd = mock_run.call_args[0][0]
+
+        # Check preset is ultrafast
+        preset_idx = cmd.index("-preset")
+        assert cmd[preset_idx + 1] == "ultrafast", f"Expected ultrafast, got {cmd[preset_idx + 1]}"
+
+        # Check crf is 28
+        crf_idx = cmd.index("-crf")
+        assert cmd[crf_idx + 1] == "28", f"Expected crf 28, got {cmd[crf_idx + 1]}"
+
+    @patch("video_processing.unified_render_service.run_ffmpeg")
+    def test_execute_ffmpeg_normal_uses_medium(self, mock_run):
+        from video_processing.unified_render_service import (
+            RenderLayer,
+            UnifiedRenderService,
+        )
+
+        plan = MagicMock()
+        plan.id = "test_plan"
+        plan.config = {}
+
+        clip = self._make_clip()
+
+        svc = UnifiedRenderService(
+            plan=plan,
+            clips=[clip],
+            asset_path_map={"a1": Path("/tmp/fake.mp4")},
+            work_dir=Path(tempfile.mkdtemp()),
+            output_width=1280,
+            output_height=720,
+            is_preview=False,
+        )
+
+        layers = [RenderLayer(role="main", clips=[clip])]
+        filter_complex, input_args = svc._build_filter_complex(layers)
+        output_path = Path(tempfile.mkdtemp()) / "out.mp4"
+        svc._execute_ffmpeg(filter_complex, input_args, output_path)
+
+        mock_run.assert_called_once()
+        cmd = mock_run.call_args[0][0]
+
+        preset_idx = cmd.index("-preset")
+        assert cmd[preset_idx + 1] == "medium"
+
+        crf_idx = cmd.index("-crf")
+        assert cmd[crf_idx + 1] == "23"
+
+
+# ── 3. RenderAdapter passes is_preview ──
+
+
+class TestRenderAdapterPreviewPassthrough:
+    """RenderAdapter 正确传递 is_preview 参数。"""
+
+    def test_render_from_memory_passes_is_preview(self):
+        from video_processing.render_adapter import RenderAdapter
+
+        db = MagicMock()
+        adapter = RenderAdapter(db)
+
+        plan = MagicMock()
+        plan.id = "test_plan"
+        plan.config = {"export": {"resolution": "854x480"}}
+
+        clip = MagicMock()
+        clip.id = "c1"
+
+        with patch.object(adapter, "_do_render") as mock_do_render:
+            mock_do_render.return_value = MagicMock(
+                success=True,
+                output_path=Path("/tmp/out.mp4"),
+                thumbnail_url="",
+                duration=5.0,
+                file_size=1000,
+                width=854,
+                height=480,
+                output_url="https://oss/test.mp4",
+                rendered_clip_ids=["c1"],
+                failed_clip_ids=[],
+            )
+
+            adapter.render_from_memory(
+                plan=plan,
+                clips=[clip],
+                asset_path_map={"a1": Path("/tmp/fake.mp4")},
+                is_preview=True,
+            )
+
+            mock_do_render.assert_called_once()
+            _, kwargs = mock_do_render.call_args
+            assert kwargs.get("is_preview") is True
+
+
+# ── 4. 预览模式跳过 ASR ──
+
+
+class TestPreviewSkipsASR:
+    """预览模式跳过 ASR 初始化。"""
+
+    def test_render_method_source_has_asr_skip(self):
+        """_do_render 在 is_preview=True 时不调用 _get_asr_service。"""
+        with open("apps/worker/video_processing/render_adapter.py") as f:
+            source = f.read()
+
+        assert (
+            "None if is_preview else self._get_asr_service()" in source
+        ), "Should skip ASR initialization in preview mode"
+
+
+# ── 5. 并行下载逻辑 ──
+
+
+class TestParallelDownload:
+    """generation.py 并行下载素材。"""
+
+    def test_parallel_download_uses_thread_pool(self):
+        with open("apps/worker/worker_app/tasks/generation.py") as f:
+            source = f.read()
+
+        assert "ThreadPoolExecutor" in source, "Should use ThreadPoolExecutor for parallel downloads"
+        assert "as_completed" in source, "Should use as_completed for result collection"
+
+    def test_parallel_download_preserves_order(self):
+        with open("apps/worker/worker_app/tasks/generation.py") as f:
+            source = f.read()
+
+        assert "sorted(results_map.keys())" in source, "Should sort results by original index"
+
+
+# ── 6. generation.py _render_video passes is_preview ──
+
+
+class TestRenderVideoPassesPreview:
+    """_render_video 正确传递 is_preview 到 render_from_memory。"""
+
+    def test_render_video_passes_is_preview(self):
+        with open("apps/worker/worker_app/tasks/generation.py") as f:
+            source = f.read()
+
+        assert "is_preview=is_preview" in source, "Should pass is_preview to render_from_memory"
+
+
+# ── 7. Preview mode skips thumbnail and validation ──
+
+
+class TestPreviewSkipsThumbnailAndValidation:
+    """预览模式跳过缩略图生成和输出校验。"""
+
+    def test_render_adapter_skips_thumbnail_in_preview(self):
+        with open("apps/worker/video_processing/render_adapter.py") as f:
+            source = f.read()
+
+        assert "if not is_preview:" in source, "Thumbnail should be conditional on is_preview"
+
+    def test_render_adapter_skips_validation_in_preview(self):
+        with open("apps/worker/video_processing/render_adapter.py") as f:
+            source = f.read()
+
+        assert "预览模式：跳过输出校验" in source, "Should skip validation in preview mode"
+
+
+# ── 8. Pass-through rendering uses ultrafast in preview ──
+
+
+class TestPassThroughPreviewPreset:
+    """直通渲染在预览模式也使用 ultrafast。"""
+
+    def test_pass_through_has_preview_preset(self):
+        with open("apps/worker/video_processing/unified_render_service.py") as f:
+            source = f.read()
+
+        # The pass_through method should also use ultrafast for preview
+        # Count occurrences of "ultrafast" - should be at least 2 (execute_ffmpeg + pass_through)
+        count = source.count('"ultrafast" if self.is_preview')
+        assert count >= 2, f"Expected at least 2 ultrafast preset usages, found {count}"