The process of SVM algorithm finding support vector involves the calculation of n-order matrix. N is the number of samples. When the number of samples is large, the calculation of high-order matrix will consume a lot of computing time. At the same time, the performance of SVM model depends on the setting of penalty parameters and kernel parameters, The traditional cycle verification method of parameter optimization has high time complexity. In order to solve the above two problems, this paper proposes a high-performance SVM model training algorithm by using convex hull algorithm to sparse the training samples, and optimizing the selection of penalty parameters and kernel parameters through genetic algorithm.