The integrated electronic-pneumatic shift system (EPSS) represents the current trend in the field of automated manual transmissions (AMT) for heavy-duty commercial vehicles. However, due to the switching characteristics of the pneumatic on/off solenoid valves, there are mutual coupling mechanisms and constraints between continuous state variables and discrete events in the gear shifting process, so that the control approach designed based on the continuous dynamics model is difficult to directly control the on/off solenoid valves with discrete dynamic characteristics. To address this issue, this paper proposes a hybrid model predictive control (HMPC) approach for achieving multi-objective optimization, specifically precise controlling of actuator displacement during gear shifts while reducing solenoid valves switching frequencies. Firstly, the mixed behavior due to the continuous dynamics of the shift actuator displacement and the discrete characteristics of the solenoid valves is modeled. Segmental linear approximations are made to the pneumatic nonlinear characteristics of the system. Both auxiliary logic variables and continuous logic variables are introduced, and a mixed logic dynamics (MLD) model is constructed in a hybrid system description language. Secondly, the shift actuator displacement control strategy is designed based on the mixed model predictive control theory. The displacement trajectory control problem is transformed into a constrained finite time domain optimization problem by establishing an objective function that reflects the actual control requirements of the system. It is further transformed into a mixed-integer quadratic programming (MIQP) problem for solution to achieve an effective synthesis of the mixed-model predictive control law for the displacement trajectory of the shift actuator. Finally, the designed mixed-model predictive controller is experimentally verified. The results show that the approach can achieve accurate displacement trajectory control and lower solenoid switching frequency compared with MPC and PID controllers.