The global business circumstances are rapidly changed so the design optimization which is related to quality engineering is needed than ever. In case of the robot casting the tolerance of material and variation of material property are the main cause of quality issues and such a reason quality based design optimization is urgently needed. Robust design gives robustness to the objective function considering tolerance and variation of material property while in the reliability based design, they are controlled by probability of constraint violation. In this study, two other quality engineering methods are done about same robot casting and compare with each other. During that process, some technical theory is applied. Sensitivity analysis with the design of experiments is used for reducing control design variable, approximation with central composite design is used to make objective and constraint functions. Genetic algorithm and search direction method is applied to make robust design optimization. For reliability based design optimization, the performance measure approach method is used. Comparing two other quality engineering method means that the robustness of objective and constraint functions which is resulted from robust design has certain results in aspect of reliability which is controlled by probability of constraint violation.