Abstract: The hydraulic accumulator ( SOC) load model of hydraulic hybrid vehicles based on neural network control is established. By selecting a certain heavy-duty hydraulic hybrid vehicle , the simulation model under Matlab / Simulink platform is built. In the FUDS simulation cycle ，the actual vehicle speed and expected speed, vchicle speed error parameters，the SOC predict state parameters，etc. are analyzed. lt is concluded that when the opening value of hydraulic accumulator ACC is 0.2 and the closing value is 0. 5 ,the coordination between engine and accumulator is the best. 'The response of SOC valuc is good,and the expected effect is achieved.

Key words: hydraulic hybrid power, FUDS, neural network control hydraulic accumulator, state of charge, speed error