Abstract: As coal mining continues to extend to deeper levels, the working face faces increasing geological pressure, the risk of gas release and accumulation significantly increases.Additionally, the physical properties and structural characteristics of coal seams in deep mines differ from those in shallow areas, further increasing the potential risk of coal and gas outbursts.In this study, based on data from a specific mine, the Boxplot and Multiple Imputation(MI)methods were applied for data cleaning.The influential factors were selected through correlation analysis, and a prediction index system for coal and gas outburst was established using the Boxplot-MI-C approach.Subsequently, a model framework based on Convolutional Neural Network(CNN)in deep learning was developed.The model hyperparameters were optimized using the Cormorant Optimization Algorithm(COA).This resulted in the establishment of a COA-CNN model for predicting coal and gas outburst.Support Vector Machine(SVM),COA-SVM,Artificial Neural Network(ANN),COA-ANN and CNN models were established for comparative validation.The results demonstrated that the SSA-CNN model achieved the highest accuracy in predicting outcomes, exhibiting superior robustness and generalization capability.This model can provide better decision-making references for the forecast and and control of coal and gas outburst disaster.