Aiming at the problems of difficult control of electric traction equipment for heavy haul trains and low accuracy of control models, a distributed fuzzy sliding mode optimization control strategy for heavy haul trains based on a predictive compensation model is proposed. Firstly, based on the complex road condition information of heavy haul railways, a longitudinal dynamics model of heavy haul trains is established. Then, a long short-term memory neural network (LSTM) optimized by the Northern Goshawk Optimization (NGO) algorithm is used to predict and compensate the uncertain terms in the model. At the same time, a variable universe fuzzy algorithm (VUF) is used to improve the fast terminal sliding mode controller (FTSM). Finally, the improved controller is combined with the air braking strategy to jointly control the corrected model. The simulation results based on the heavy haul train operation simulation platform and real operation data show that the proposed method can effectively improve the handling stability and safety of 10,000-ton heavy haul trains.