Abstract: Dynamic discrimination method of gas security situation of stope face and six kinds of diversified evaluation indexes were proposed started from two major gas discharge forms of gas drainage and underground ventilation, and combining three major methane sources of adjacent layers gas emission, coal tunnel wall gas emission, and gas emission caused by coal drop. Dynamic discrimination method of gas security situation of stope face based on BP neural network was put forward too. Matlab validation results show that the method will follow the index to find the most optimized sequence when it discriminates relative changes of gas security situation of stope face, on the basis of given indicators, the method can accurately discriminates relative changes of gas security situation, but absolute discrimination results need to be judged specifically combined with sub-index values.