In this paper the multi-objective optimization design of a preliminarily designed end-effector tool for robot-assisted automatic polishing system is presented. ADAMS and iSIGHT are integrated for the multi-objective optimization design, the overall procedure of which is shown in this paper. Design of Experiment is arranged before optimization. The global optimization methods provided by iSIGHT software- Exterior Penalty and Multi-Island Genetic are utilized to perform optimizing process. Simulation executed in ADAMS with model updating with optimized structural design parameters verifies the significance of the multi-objective optimization in improving the performance of preliminarily designed end-effector tool.