diff --git a/astrbot/core/provider/sources/mimo_tts_api_source.py b/astrbot/core/provider/sources/mimo_tts_api_source.py
index 2966bfb7d..63389a10b 100644
--- a/astrbot/core/provider/sources/mimo_tts_api_source.py
+++ b/astrbot/core/provider/sources/mimo_tts_api_source.py
@@ -88,13 +88,15 @@ class ProviderMiMoTTSAPI(TTSProvider):
             }
         )
 
+        audio_params = {"format": self.audio_format}
+        # voice design 模型不支持 audio.voice 参数
+        if "voicedesign" not in self.model_name:
+            audio_params["voice"] = self.voice
+
         return {
             "model": self.model_name,
             "messages": messages,
-            "audio": {
-                "format": self.audio_format,
-                "voice": self.voice,
-            },
+            "audio": audio_params,
         }
 
     async def get_audio(self, text: str) -> str:
diff --git a/tests/test_mimo_api_sources.py b/tests/test_mimo_api_sources.py
index c2b02aa13..ec5b67453 100644
--- a/tests/test_mimo_api_sources.py
+++ b/tests/test_mimo_api_sources.py
@@ -129,6 +129,39 @@ def test_mimo_tts_seed_text_is_not_prepended_to_assistant_content():
         asyncio.run(provider.terminate())
 
 
+def test_mimo_tts_voicedesign_model_omits_voice_param():
+    """voice design 模型不应包含 audio.voice 参数"""
+    provider = _make_tts_provider(
+        {
+            "model": "mimo-v2.5-tts-voicedesign",
+            "mimo-tts-seed-text": "",
+        }
+    )
+    try:
+        payload = provider._build_payload("hello")
+        assert "voice" not in payload["audio"]
+        assert payload["audio"]["format"] == "wav"
+    finally:
+        asyncio.run(provider.terminate())
+
+
+def test_mimo_tts_regular_model_includes_voice_param():
+    """普通 TTS 模型应包含 audio.voice 参数"""
+    provider = _make_tts_provider(
+        {
+            "model": "mimo-v2.5-tts",
+            "mimo-tts-voice": "custom_voice",
+            "mimo-tts-seed-text": "",
+        }
+    )
+    try:
+        payload = provider._build_payload("hello")
+        assert payload["audio"]["voice"] == "custom_voice"
+        assert payload["audio"]["format"] == "wav"
+    finally:
+        asyncio.run(provider.terminate())
+
+
 def test_mimo_headers_use_single_authorization_method():
     assert build_headers("test-key") == {
         "Content-Type": "application/json",