Abstract: This paper uses ADAMS software to establish a multi-body model of a semi-active suspension system. The suspension model is imported into MATLAB/Simulink from ADAMS/View by ADAMS/Control module, and then, a fuzzy logic controller is completed in MATLAB/Simulink environment. The fuzzy rules of controller are optimized by an improved genetic algorithm written in Matlab language, and that accomplishes the optimization design for the semi-active suspension systems based on improved genetic algorithms. In order to test the control efforts and the performance of the improved genetic algorithm, the semi-active suspension system is analyzed with random road profile at two different speeds, 25m/s and 35m/s, by the comparison with the passive suspension. The results show that the fuzzy logic controller based on improved genetic algorithm enhances the ride comfort performance of the vehicle significantly.