Abstract: Aiming at the catastrophic problem of common coal mine gas explosion, this paper combines the trend surface analysis method to establish a gas explosion warning model, and validates the model through actual data, and the model has good robustness.Through the data fusion of the comprehensive information of gas explosion, the gas and oxygen concentration are selected as the key indicators of the early warning model, and the trend surface analysis of the finite discrete data is established.The trend surface warning model of gas and oxygen concentration is established, and the distribution of key indicators is obtained.Features that enable predictions and warnings of potential gas hazards.