The invention provides an audio processing system and method based on echo suppression switch in voice detection. The system comprises a local voice detection module, a network voice detection module,an attenuator module, a switch module, an echo suppression module, a speaker and a pickup device. When the local voice detection module determines a local voice, and the network voice detection module determines no voice in the network through a voice detection method, the switch module forwards an audio stream C not passing through the echo suppression module to an audio stream E, so as to reduce distortion of the audio stream E, the attenuator module is started to attenuate an audio stream A, so as to prevent the audio stream A, the background noise, from inhibiting the network to send theaudio stream E, further, slight background sound of the audio stream A received by the network is kept. Through adoption of the system and method, the local voice processing process eliminates necessary echo and reduces echo suppression, so that the voice is less damaged, and the sound quality of the network audio stream E sent locally is improved.