The energy management strategy (EMS) is a critical component of the design of hybrid electric vehicles. In this paper, a predictive EMS based on Nonlinear Model Predictive Control (NMPC) and power demand estimation approaches is proposed for the EMS of battery and diesel engine in a hybrid train. The Artificial Neural Network (ANN) method is introduced to realize model prediction of power demand. On this basis, the NMPC-based control framework is constructed, and a nonlinear programming approach with excellent global search ability and faster convergence is used to find the optimal control sequence in the receding horizon. Theoretical simulations are performed and the findings show that the total cost of EMS with implemented PWM as an AC/DC converter reduces diesel engine power consumption rather than using a conventional rectifier at 34.1%. © 2023 IEEE.