Abstract: In order to accurately predict the amount of mine gas emission, reduce the harm caused by gas emission, through gray correlation analysis theory, we indicate that the main factors affecting gas emission is original gas content > thickness of coal seam > depth of coal seam > length of working surface > advancing speed > inclination of coal seam. The prediction model of gas emission is constructed by RBF model optimization, and we use Matlab simulation to predict the mine gas emission. The results show that simulation prediction result based on the RBF model optimization is very close to actual gas emission, in 5 sets of forecast data, the maximum error is 3.6%, the minimum error is 0.8%, the average error is 1.84%. The prediction accuracy is higher and can be applied to the prediction of mine gas emission.