Abstract: Based on the algorithm of classification and regression tree (CART (Classification And Regression Tree)),a discriminate model of coal and gas outburst was established according to main factors,which are gas pressure,gas initial velocity,destructive type of coal,hard coefficient and depth.Select data of discriminating indexes from typical mines with coal and gas outburst as training data,and find the champion tree by the method of k - fold cross-validated,then extract the optimized rules,at lased,the rules is used to discriminate the coal and gas outburst in the instance.The results show that CART model is a simple and effective method,it can be used in discriminate analysis of coal and gas outburst as a auxiliary method.