Abstract: Combining the reliability theory with optimization technique, the reliability-based optimization design for vehicle parts is emphatically discussed, then a numerical method is proposed for the reliability-based optimization design. Based on the premise that the probability distribution of random parameters has been known, the probability density function and probability distribution function of random structural response under external loading are accurately and quickly obtained by way of saddlepoint approximation and, as a result, the saddlepoint approximation method is proved accurate with higher computing speed in comparison with the Monte-Carlo method. Therefore, the reliability of vehicle parts computed by the saddlepoint approximation lays a theoretical foundation for the reliability-based optimization of vehicle parts, and it can accurately and quickly acquire the design information.