On account of the shortcomings of particle swarm optimization algorithm(PSO), such as poor global search capability and low convergence accuracy, levy flight [1] and Gaussian mutation [2] are used to propose particle swarm optimization algorithm based on levy flight. In iteration, the particle aggregation degree [3] is calculated, and the corresponding probability is selected to carry out levy flight [4] according to the particle aggregation degree, it strengthen the global optimization ability. Gaussian variation is carried out on particle positions of each iteration, and particles with better fitness are selected for iteration, which enhances the local search capability. At the same time, adaptive perturbation is carried out to the global optimal position of each iteration to increase the optimization precision of the optimal value. The test shows that the convergence rate of the improved algorithm is better than that of PSO, and the convergence accuracy is higher.