To improve vehicle ride comfort and handling stability, a multi-objective optimization strategy based on improved genetic algorithm was designed for the suspension system. As the study of a light passenger bus, taken the vehicle body side angle, yaw rate and the amplitude of vibration acceleration as optimization index, a multi-objective optimization model of suspension system parameters was established. The method improved the selection mechanism in the individual populations, and the improvements would enable the parent population and offspring population of optimal individuals to have the same opportunity to be selected to ensure diversity of new species. The real vehicle experimental results showed that: compared with the suspension before optimization, the vehicle body side angle, yaw rate and the amplitude of vibration acceleration had been reduced by 12.3%, 6.4% and 9.8%. The new strategy could reasonably match the parameters of suspension system, and simultaneously improve the vehicle ride comfort and handling stability.