Abstract: Combined with characteristics of environment information in coal mine underground,the paper introduced a risk assessment system of environment information in coal mine underground based on multi-sensor information fusion(MSF).The system establishes model of three-layer error back propagation neural network used in risk prediction of environment information in coal mine underground and adopts information fusion algorithm of neural network to analyze and treat sample data.The simulation results showed that the system can assess risk degree of environment information in coal mine underground more accurately and has better robustness and generalization ability.