Abstract: In order to improve the efficiency and accuracy of coal seam outburst risk assessment, Distance Discriminant Analysis (DDA) was introduced, five indexes including initial speed of methane diffusion, the coefficient of solidity of coal, coal seam gas pressure, type of coal body failure and mining depth, were regarded as discriminant factors. Grey Relational Analysis (GRA) was used to calculate the weight matrices, a weighted distance coupled discriminant model for coal seam outburst risk assessment was established.30 examples of coal and gas outburst were trained as learning samples, corresponding discriminant criteria was set up, and the ratio of mistake-distinguish was zero after training. The GRA-DDA model was used to evaluate 10 coal and gas outburst cases, and the evaluation results of this model was verified by comparing with the single index method, BP neural network method and unweighted distance discriminant analysis method. The results showed that the evaluation results of the GRA-DDA model are accurate and accord with the actual situation.