Abstract: Mine gas emission prediction of coal mine safety production is very important work, In view of the advantages of principal component analysis and stepwise regression analysis, this paper combines two methods to establish a gas emission generation regression prediction model. Taking the No. 30 coal seam of Junde Coal Mine as an example, four principal component factors affecting the gas emission from the mining face were obtained by principal component analysis, and the stepwise linear regression analysis was used to predict the gas emission from the mining face. The results show that the principal component-stepwise regression analysis reduces the number of variables that need to be considered in the regression analysis. The prediction results have better accuracy, and the prediction accuracy is better than the one-way regression prediction and multiple regression prediction. It has a good application prospect.