The output of MEMS gyroscope is affected by temperature, which will produce temperature drift and affects the precision of gyroscope. To solve this problem, this paper proposes a method based on GA optimized support vector machine to compensate the temperature of gyroscope. Seven temperature points are selected in the whole temperature region as samples. Support vector machine (SVM) based on radial basis kernel function is adopted for modeling, and GA is used to optimize SVM parameters to further improve the accuracy of gyro temperature compensation. Experimental results show that the method proposed in this paper is more effective than traditional least square method and BP neural network. © 2019 IEEE.