In order to effectively avoid the problem of imbalance between production and sales of new energy vehicles, the PSO algorithm is used to optimize the parameter iteration process of the BP neural network, to make up for the shortcomings of optimizing the original BP neural network which is easy to fall into the local optimum and slow convergence, and to construct a new energy vehicle sales prediction model based on the PSO-BP neural network, and to take BYD as an example to carry out the prediction of the exponential smoothing method, the BP and PSO-BP neural network. The prediction of BYD, BP and PSO-BP neural network are carried out. The results show that the BP neural network model has a significant advantage over the exponential smoothing model in terms of MSE, MAE and MAPE indicators, and the BP neural network model optimized by the particle swarm algorithm shows a decrease of nearly 7×10^7 in MSE, 3 346 in MAE and 1.71% in MAPE. It can be seen that the new energy vehicle sales prediction model based on PSO-BP neural network is better than the exponential smoothing model and BP neural network model, particle swarm optimized BP neural network can make the model jump out of the local optimum, accelerate the convergence speed, the prediction results of the error rate is lower, the accuracy is higher, and it has a guiding role in the planning and production of enterprises.