The relationship between customer demand and the modular instances of configurable product was established to solve the product configuration optimization problem. Constrained terms were processed by the dynamic penalty function and the mathematical model was built. The basic PSO algorithm was improved by using crossover and mutation strategy of genetic algorithm and Metropolis rule of simulated annealing algorithm. The hybrid PSO algorithm was used to solve modular product configuration optimization, and the binary encoding scheme was adopted to express the configurable product cases. Finally, a simulation comparison of product configuration optimization was carried out by the hybrid PSO and genetic algorithm respectively. The results show that the hybrid PSO algorithm is simple and effective.