In order to improve the ride comfort of a domestic self-dumping truck, a virtual prototype vehicle model was built through multi-body dynamics software Adams. And vehicle road test of ride comfort was implemented to verify the validity of this model. Taking the suspension stiffness and damp parameters of cab suspension as design variables, the root mean square (RMS) values of floor’s vertical and seat’s pitching weighted acceleration as optimization objectives, and the deflection of front and rear cab suspension as constraints, the Kriging approximation models were constructed based on optimal Latin hypercube design. On this basis, multi-objective optimization for self-dumping truck ride comfort was performed with particle swarm optimization algorithm, and the Pareto optimal set was obtained. Furthermore, a vehicle road test of ride comfort was conducted by using one of the optimal solutions. The results indicated that the Kriging approximation model with high fitting accuracy could significantly improve the efficiency of ride comfort optimization of self-dumping truck according to different weighting schemes of optimization objectives. The road test results showed that the overall weighted acceleration RMS values on the cab floor of the improved self-dumping truck were greatly reduced with a maximal reduction of 16.5%.