Abstract: Emergency evacuation of underground personnel is an important safety measure when a water disaster occurs in a mine. In this paper,according to the demand of the simulation for emergency evacuation and the analysis of vari- ous complex influencing factors,the geological environment and the underground scene of the mining area are construc- ted based on the data acquisition and identification of the Internet of Things. The simulation and visualization of emer- gency evacuation in VR simulation environment is carried out through the information acquisition of water bursting points,the simulation of roadway flow and behavior of evacuees and intelligent dynamic path search. This method achieves a dynamic process simulation of individual or group behavior as well as generates safe and rapid evacuation routes. Qualitative and quantitative evaluation and early warning are conducted for emergency evacuation decisions and evacuation exercises. The method has been used in mine,and the results show that it can provide an auxiliary decision- making platform for personnel evacuation plan,emergency rescue and personnel training under the situation of mine disaster.