In order to optimize the ride comfort of mining dump trucks of 393 tons in bad driving traffic, the nonlinear stiffness and damping characteristics of hydro-pneumatic suspension were optimized. A 7-DOF full model, more realistic than traditional vehicle models, was derived. And considering the uncertainty of mass and position of centroid in the process of transportation, some factors of stiffness and damping of suspension having greater influence on target were selected as the design variables. The target function contains the body's vertical acceleration, roll angular acceleration, pitch angular acceleration, dynamic displacement of suspension and dynamic load of the wheel. An approximate model between design variables and objective function was established by High Dimensional Model Representation based on Thin Plate Spline Interpolation (TPS-HDMR).This model adopts multi-objective optimization with genetic algorithm based on the concept of Pareto, and then gets the Pareto optimal solution set. The simulation results have shown that multi-objective optimization can improve the performance of 11 targets and ride comfort.