Abstract: In order to improve the ionospheric critical frequency based on BP neural network (f_oF₂) prediction accuracy, we adopt a method of improved particle swarm optimization neural network method to optimize the initial weights of BP network to prevent the emergence of local optimal neural network in training. By comparing the results of neural network prediction based on particle swarm optimization and the optimization of genetic algorithm, we find that the two methods have good performance for BP neural network. In addition, compared with the international reference ionosphere model (IRI2016), we find that the adaptive mutation particle swarm optimization neural network can effectively improve the prediction accuracy of f_oF₂, and has better prediction effect in low latitude area.