In order to optimize the arithmetic of the automobile suspension springs, the optimization algorithm of spring stress model based on BP neural networks was proposed. The various models of suspension spring were aggregated and analyzed, the parallel genetic algorithm for the suspension springs was proposed in this paper as well. And the spring models can be optimized fast and efficiently by the algorithm. Aiming at different models, the maximum stress point and the distribution law are given by means of experimental verifications. A new stress analysis method and a new estimate rule for the design of automotive suspension springs are provided by this method.