Abstract: A plan of a class of multiple models system was put forward, in which there is a kind of intelligent hierarchical supervisory control system(IHSC) that is a supervisor to control the engine subsystems of the multiple energy source assembly control system of the hybrid electric vehicle(HEV). A probability neural network(PNN) made the organizing level of IHSC, which equals to the identification system of HEV processes and indicates the working procedure of HEV, the hidden level function was made of Parzen window kernel function, the identification result depended on the biggest probability value of the output level of PNN. A fuzzy decision-maker, which belongs to the multiple attributes decision and optimizes the working rate between the motor system and engine system according to the results of the organizing level of IHSC, made the coordination level of IHSC. A simulation study of the multiple energy source assembly control system of HEV was given to verify its feasibility.