Abstract: In order to solve the problems of low accuracy and slow response speed of existing support vector machine (SVM)-based coal and gas outburst prediction methods, a risk prediction method of coal and gas outburst based on improved grey wolf optimizer (IGWO) optimized SVM is proposed. The influence degree of each influencing factor on coal and gas outburst is analyzed by using the grey relational entropy weight method, and gas pressure, gas content, initial gas diffusion speed and mining depth are extracted as main control factors of coal and gas outburst according to the correlation degree order, and the main control factors are divided into a training set and a test set, and normalized. In order to improve the defects of the traditional grey wolf optimizer (GWO) population easily falling into local optimum and slow optimization speed, the out-of-bounds processing mechanism and the random difference mutation strategy embedded in Levy flight are introduced to improve the grey wolf optimizer (ie IGWO), so as to improve the convergence precision and speed of GWO effectively. The core parameters and penalty parameters of SVM are optimized by IGWO, and the main control factors of coal and gas outburst are input into IGWO-SVM for classification. And the classification results are compared with the actual test set so as to realize the risk prediction of coal and gas outburst. The simulation results show that compared with the prediction methods based on whale optimization algorithm-SVM ( WOA-SVM), grey wolf optimizer-SVM ( GWO-SVM) and particle swarm optimization-SVM ( PSO-SVM), the prediction method based on IGWO-SVM has higher prediction precision, and can meet the precision and reliability requirements of coal and gas outburst prediction while improving the operation efficiency of SVM. The accuracy rate reaches 96.67% and the prediction speed is 5.58 s.