Abstract: Based on the on-site audio and video information of emergency rescue and emergency handling in mining area, this study uses H264 protocol to realize video encoding and decoding, and uses Speex voice engine to realize audio encoding and decoding. Relying on wireless Mesh network, combined with RTP real-time transmission protocol and SIP protocol based voice intercom technology, the visual on-site command system of emergency rescue based on audio and video is realized. The system transmits the scene images, environmental parameters, vital signs of rescue team members and other information to the backstage command center through camera, hearing device, vital signs sensor and multi-parameter gas detector. At the same time, the system supports two-way voice intercom between the front and the rear, so that the rescue commanders can grasp the whole rescue situation in real time and carry out visual management and scheduling. Through the emergency rescue practice of several rescue teams in the domestic coal industry, it shows that the system improves the safety, reliability and flexibility of disaster relief work, provides an effective way to prevent the occurrence of secondary disasters, and achieves good rescue effect.