Abstract: In order to meet the demand of dynamic monitoring of coal mine gas concentration, an optimization scheme based on multi-point gas detection technology is proposed in this paper. The optimized scheme uses high-precision MEMS gas sensor array to improve the spatial resolution of detection and the quality of data acquisition. Then, the improved LSTM deep learning algorithm is used to construct the prediction model of gas concentration, which realizes the high precision prediction of gas concentration. Finally, through the organic combination of these two technologies, a set of high-precision, real-time and reliable multi-point gas detection system is constructed.