Abstract: Due to the complex geology of coal seam, the prediction accuracy of gas emission is limited. The prediction of pre-pumped coal seam gas emission based on the Stacking integrated model is put forward. The normalization processing technology is used to pretreat the coal seam data from the pre-pumped coal seam drilling along the strata. Statistical analysis and machine learning algorithm are used to filter the eigenvalues of the processed data. A prediction model based on the Stacking integrated model is constructed to predict the gas emission in pre-pumped coal seam from drilling along the bedding. The experimental results show that compared with the traditional forecasting method, this method shows higher prediction accuracy and wider applicability.