In order to reduce lateral vibration of high-speed strain, the dynamics model for semi-active suspensions of vehicles was established in ADAMS/Rail code. Programs for genetic algorithm procedures were written in MATLAB code using a mixed encoding method of float and decimal numbers, and the method of adaptive genetic crossover and mutation probability based on standard deviation of fitness values. The membership functions, fuzzy rules, quantification factors and scaling factors were optimized. Using the RMS of lateral vibration at the front and the rear of the vehicle’s body as the performance index of the genetic algorithm, the fuzzy controller of the semi-active suspension was constantly optimized. The joint simulation results show that using GA-optimized fuzzy semi-active suspension can effectively reduce the lateral vibrations and improve the ride comfort of vehicles.