This study presents a novel optimization approach for high-speed vehicle positioning, leveraging BP neural networks and beamforming in the 5G communication context. Using BP neural network to predict the speed data of vehicle movement in the context of 5G mobile communication system, which is estimated by Doppler. Optimize beamforming design through this method to maximize its gain. Due to the signal attenuation of traditional GPS during high-speed driving, the positioning accuracy may decrease. Therefore, by combining beamforming technology with Doppler frequency shift measurement, the estimation of vehicle speed and position can be made more accurate. This method solves the limitations of traditional GPS under high-speed conditions. During the research process, it was found that a slightly higher signal-to-noise ratio (SNR) can greatly reduce the error in estimating speed, thereby making the positioning system more stable and accurate. The simulation results show that this method is very useful in high-speed moving scenes, and can provide a powerful solution for autonomous vehicle, traffic management and logistics tracking.