The invention relates to the field of vehicle control, in particular to a nonlinear model prediction control method of regenerative braking of an electric vehicle. The method adopts a nonlinear model prediction control strategy of a rapid system, on the basis of establishing an electric vehicle regenerative braking nonlinear model, subsystem operating state signals, acquired by sensors, of the entire vehicle, wheels, suspensions and motors in the braking process of the electric vehicle are acquired, and multiple purposes, namely controlling speed accurately, guaranteeing vehicle stability, improving driving smoothness and regenerative braking energy recovering efficiency, in the braking process are controlled coordinately. The control system is mainly applied into the driving and braking processes of traditional vehicles and electric vehicles, especially the multipurpose controlling process of strong nonlinear systems.