This paper, based on uniform design and quantum-behaved particle swarm optimization (QPSO), aims to solve the sensitivity problem of a railway vehicle system with fourteen degrees of freedom and nonlinear coupled differential equations of motion. The vehicle is very sensitive to wheel conicity if suspension parameters are not designed properly. A new heuristic nonlinear creep model is applied and is determined by adding the linear creep moment and the semi-axis lengths in the nonlinearity of the saturation constant. The vehicle's critical hunting speed is determined by Lyapunov's indirect method. This paper has demonstrated that the integration of uniform design and quantum-behaved particle swarm optimization can effectively determine the optimal solutions for suspension parameters. After applying the optimization procedure presented in this paper, the vehicle's critical hunting speed can be no longer sensitive to wheel conicity.