Abstract: The traditional slope parking is realized based on the friction force of brake system. There is a possible impact or backsliding risk because of the delay or advance of parking braking force release during start process. To solve those problems, a slope motor-driven still hold (MSH) system was proposed for electric vehicles. A dynamic model of MSH system was established based on the parameters of the driving motor system. For the uncertainty of slope and vehicle mass parameters in MSH system, a parameter identification-disturbance rejection control strategy was proposed. The parameter identification was used as the initial input value of the controller, and the disturbance rejection control technology was used to compensate the identification error. Experimental results demonstrate that, the variable forgetting factor least square method can estimate the slope and vehicle mass parameters, and the error is within 15%. In addition, the active disturbance rejection control (ADRC) can eliminate those error and disturbance. Compared with the traditional PI controller, it possesses the advantages of short backward displacement and rapid response speed.