Active steering control can significantly improve the curving performance of rail vehicles. However, it is constrained by the traction and braking of bogie. In this study, methods for optimizing and distributing active steering control quantities under constrained conditions are investigated. First, the wear number of the wheelset is set as the objective function. A controller is developed in the control layer based on a linearized bogie model with the wear number as the cost function using the model predictive control method. Then, the system state is observed using a Kalman filter with actuator feedback; the wheelset force state is also estimated using the actuator feedback. Subsequently, the optimum wheelset yaw angle under the constraints is solved. In the distribution layer, the spread of actuator displacements under traction and braking conditions is achieved by combining the longitudinal force coordination equations for the wheelset. The simulation results showed that the full utilization of the overdrive characteristics of the two coaxial actuators enabled the redistribution of actuator displacement commands under traction and braking conditions without reducing steering efficiency. Moreover, the maximum load on the actuators was effectively reduced. Finally, the parametric robustness of the controller was examined. The results showed that the controller and observer had good parametric robustness owing to the attitude stability of the bogie after active steering control was implemented.