In order to improve the accuracy of accelerated life model, an optimization method of accelerated life model under multi-stress condition was proposed. Considering the single stress and stress coupling factors in the working environment of the product, the general form of the multi-stress acceleration model is constructed based on the generalized log-linear model, and the multi-stress coupling matrix and multi-factor variance analysis are used to optimize the model. The K-S test is used to determine the life distribution of the product, and the improved particle swarm optimization algorithm is introduced to estimate the multi-parameters of the model. The example analysis shows that this method can provide an effective way for the construction of product accelerated life model under multi-stress conditions.