Aiming at a new type of urban road transportation system, namely the virtual rail train, as an example of a double articulated vehicle system with three bodies and six axles, a trajectory tracking control method is proposed with hinge forces between the vehicles, the operation deviation and the lateral stability of the vehicle system, to improve the dynamics performance of the long vehicle system. Firstly, the steering angles of all axles are determined to achieve the coordinated steering and trajectory tracking of the vehicle, where the axles A1, A4, and A6 steer in the model predictive control (MPC) theory, the axle A2 steers with the single point preview principle, and the axles A3 and A5 steer based on the extended Ackermann steering principle. Secondly, the driving torques of all wheels are determined based on the closed-loop compensation PID of the longitudinal hinge forces between the vehicles, to make the vehicle run at the desired longitudinal speed, considering the constraints of the vehicle motions, the vehicle speed condition and the optimization of hinge forces between the vehicles. Finally, based on the co-simulation of TruckSim and Simulink, the dynamic characteristics of the vehicle are studied with two conditions of double lane changing and compound curve. The results show that the present method not only achieves better tracking accuracy and lateral stability of the vehicle system, but also obviously reduces hinge forces between the vehicles, which effectively improves the dynamics performance of the long vehicle system.