A PSO-BP neural network simulation model was established with particle swarm optimization (PSO) optimizing biases and weights of back-propagation (BP) neural network. By using the PSO-BP neural network,low-pressure burning rate was simulated and calculated. When the important factor changes in a range,for instance HTPB(hydroxyl terminated polybutadiene,28%-32%),AP(ammonium perchlorate,30%-35%,weight-mean diameter 0.06-0.140 mm),GFP(catocene,0%-5%),the burning rate and pressure index of corresponding formulas were calculated and compared to corresponding experimental results. The results show that the calculation errors of the PSO-BP method are less than ±7%.