Abstract: In order to improve the accuracy of coal and gas outburst prediction.11 factors, including gas content, gas pressure and initial velocity of gas, were selected as the discriminant indicators. The outburst intensity of coal and gas was divided into four grades: no outburst, small outburst, medium outburst and large outburst.28 sets of data of Qianxibei Coal Mine in Guizhou were used as training samples to establish a PCA-Fisher discriminant analysis model for risk level prediction of coal and gas outburst. Using the remaining 6 sets of data as a prediction sample to test and apply the model, the discriminant result was compared with that of BP neural network model and Fisher discriminant model. The results showed that PCA-Fisher discriminant model has higher accuracy and reliability, it can predict the risk level of coal and gas outburst effectively.