diff --git a/apps/web/src/pages/generate/hooks/useStep5Voice.tsx b/apps/web/src/pages/generate/hooks/useStep5Voice.tsx
index 6256826b4..89fbe68e1 100755
--- a/apps/web/src/pages/generate/hooks/useStep5Voice.tsx
+++ b/apps/web/src/pages/generate/hooks/useStep5Voice.tsx
@@ -2,30 +2,30 @@
  * Step 5 配音选择 Hook
  * 封装 AI 推荐、预设音色试听、TTS 自定义合成、存为素材等逻辑
  */
-import { useState, useRef, useCallback } from "react"
-import { message } from "antd"
-import type { VoiceClone } from "@/api/voice-clone"
-import { VOICE_GENDER_ICON, CLONE_STATUS_CONFIG } from "../constants"
-import { formatDuration } from "../utils/formatDuration"
-import { useVoiceAudio } from "./step5-voice/useVoiceAudio"
-import { useVoiceRecommend } from "./step5-voice/useVoiceRecommend"
-import { useTtsSynthesis } from "./step5-voice/useTtsSynthesis"
-import { useSaveToLibrary } from "./step5-voice/useSaveToLibrary"
-import { useVoiceModeSelection } from "./step5-voice/useVoiceModeSelection"
+import { useState, useRef, useCallback, useEffect } from "react";
+import { message } from "antd";
+import type { VoiceClone } from "@/api/voice-clone";
+import { VOICE_GENDER_ICON, CLONE_STATUS_CONFIG } from "../constants";
+import { formatDuration } from "../utils/formatDuration";
+import { useVoiceAudio } from "./step5-voice/useVoiceAudio";
+import { useVoiceRecommend } from "./step5-voice/useVoiceRecommend";
+import { useTtsSynthesis } from "./step5-voice/useTtsSynthesis";
+import { useSaveToLibrary } from "./step5-voice/useSaveToLibrary";
+import { useVoiceModeSelection } from "./step5-voice/useVoiceModeSelection";
 
 interface UseStep5VoiceProps {
-  selectedVoice: string
-  onSelectedVoiceChange: (voiceId: string) => void
-  voiceMode: "preset" | "custom" | "clone"
-  onVoiceModeChange: (mode: "preset" | "custom" | "clone") => void
-  selectedClonedVoice: string
-  onSelectedClonedVoiceChange: (voiceId: string) => void
-  clonedVoices: VoiceClone[]
-  addClone: (voice: VoiceClone) => void
-  hasProcessing: boolean
-  cloneModalOpen: boolean
-  onCloneModalOpenChange: (open: boolean) => void
-  titleText: string
+  selectedVoice: string;
+  onSelectedVoiceChange: (voiceId: string) => void;
+  voiceMode: "preset" | "custom" | "clone";
+  onVoiceModeChange: (mode: "preset" | "custom" | "clone") => void;
+  selectedClonedVoice: string;
+  onSelectedClonedVoiceChange: (voiceId: string) => void;
+  clonedVoices: VoiceClone[];
+  addClone: (voice: VoiceClone) => void;
+  hasProcessing: boolean;
+  cloneModalOpen: boolean;
+  onCloneModalOpenChange: (open: boolean) => void;
+  titleText: string;
 }
 
 export function useStep5Voice({
@@ -43,7 +43,7 @@ export function useStep5Voice({
   titleText,
 }: UseStep5VoiceProps) {
   /* ── 子模块 ── */
-  const { playingVoice, toggleVoicePlay } = useVoiceAudio()
+  const { playingVoice, toggleVoicePlay } = useVoiceAudio();
 
   const {
     presetVoices,
@@ -52,7 +52,7 @@ export function useStep5Voice({
     voiceRecommendations,
     hasVoiceRecommend,
     handleVoiceRecommend,
-  } = useVoiceRecommend(titleText)
+  } = useVoiceRecommend(titleText);
 
   const {
     customVoiceText,
@@ -64,7 +64,7 @@ export function useStep5Voice({
     synthesizeMutation,
     handleSynthesizeVoice,
     resetTtsState,
-  } = useTtsSynthesis(selectedVoice)
+  } = useTtsSynthesis(selectedVoice);
 
   const {
     saveModalOpen,
@@ -80,7 +80,7 @@ export function useStep5Voice({
     handleOpenSaveModal,
     handleConfirmSave,
     handleAddTagInModal,
-  } = useSaveToLibrary(completedTtsJobId, resetTtsState)
+  } = useSaveToLibrary(completedTtsJobId, resetTtsState);
 
   const {
     handleSelectRecommendedVoice,
@@ -99,46 +99,78 @@ export function useStep5Voice({
     addClone,
     cloneModalOpen,
     onCloneModalOpenChange,
-  })
+  });
 
   /* ── 克隆音色试听播放 ── */
-  const [playingCloneVoice, setPlayingCloneVoice] = useState<string | null>(null)
-  const cloneAudioRef = useRef<HTMLAudioElement | null>(null)
+  const [playingCloneVoice, setPlayingCloneVoice] = useState<string | null>(
+    null,
+  );
+  const cloneAudioRef = useRef<HTMLAudioElement | null>(null);
+  const cloneMountedRef = useRef(true);
+
+  // 组件卸载时清理克隆音色试听的音频资源
+  useEffect(() => {
+    cloneMountedRef.current = true;
+    return () => {
+      cloneMountedRef.current = false;
+      if (cloneAudioRef.current) {
+        cloneAudioRef.current.onended = null;
+        cloneAudioRef.current.pause();
+        cloneAudioRef.current = null;
+      }
+    };
+  }, []);
 
   const handlePlayCloneSample = useCallback(
     (voice: VoiceClone) => {
       if (playingCloneVoice === voice.id) {
-        cloneAudioRef.current?.pause()
-        cloneAudioRef.current = null
-        setPlayingCloneVoice(null)
-        return
+        if (cloneAudioRef.current) {
+          cloneAudioRef.current.onended = null;
+          cloneAudioRef.current.pause();
+        }
+        cloneAudioRef.current = null;
+        setPlayingCloneVoice(null);
+        return;
+      }
+      // 暂停上一个音频并清除事件监听，避免竞态
+      if (cloneAudioRef.current) {
+        cloneAudioRef.current.onended = null;
+        cloneAudioRef.current.pause();
       }
-      cloneAudioRef.current?.pause()
       // TODO: 后端克隆音色试听接口出来后，改为调用接口获取试听音频
       // 目前使用 sample_url 字段，如果没有则提示
-      const audioUrl = voice.sample_url
+      const audioUrl = voice.sample_url;
       if (!audioUrl) {
-        message.warning("该克隆音色暂无试听音频")
-        return
+        message.warning("该克隆音色暂无试听音频");
+        return;
       }
-      const audio = new Audio(audioUrl)
-      cloneAudioRef.current = audio
-      audio.play().catch(() => {
-        message.error("播放失败，请检查网络")
-      })
+      const audio = new Audio(audioUrl);
+      cloneAudioRef.current = audio;
+      audio
+        .play()
+        .then(() => {
+          if (cloneMountedRef.current) setPlayingCloneVoice(voice.id);
+        })
+        .catch(() => {
+          if (cloneMountedRef.current) {
+            message.error("播放失败，请检查网络");
+            cloneAudioRef.current = null;
+          }
+        });
       audio.onended = () => {
-        setPlayingCloneVoice(null)
-        cloneAudioRef.current = null
-      }
-      setPlayingCloneVoice(voice.id)
+        if (cloneMountedRef.current) {
+          setPlayingCloneVoice(null);
+          cloneAudioRef.current = null;
+        }
+      };
     },
     [playingCloneVoice],
-  )
+  );
 
   const handleCloneSuccess = (voice: VoiceClone) => {
-    rawCloneSuccess(voice)
-    message.success("音色克隆成功！")
-  }
+    rawCloneSuccess(voice);
+    message.success("音色克隆成功！");
+  };
 
   return {
     // 数据
@@ -197,8 +229,7 @@ export function useStep5Voice({
     VOICE_GENDER_ICON,
     CLONE_STATUS_CONFIG,
     formatDuration,
-  }
+  };
 }
 
-export default useStep5Voice
-
+export default useStep5Voice;