The uncertainty of demand and lead time in inventory management has posed challenges for the supply chain management. The purpose of this paper is to optimize the total pro t and customer service level of supply chain by robust parameter design of inventory policies. This paper proposes system dynamics simulation, Taguchi method, and Response Surface Methodology (RSM) to model a multi-echelon supply chain. Based on the sequential experiment principle, Taguchi method combining location with dispersion modeling method is adopted to locate the optimum area quickly, which is very e cient to optimize the responses at discrete levels of parameters. Then, fractional factorial design and full factorial design are used to recognize signi cant factors. Finally, RSM is used to nd the optimal combinations of factors for pro t maximization and customer service level maximization at continuous levels of parameters. Furthermore, a discussion of multiresponse optimization is addressed with di erent weights of each response. Con rmation experiment results showed the e ectiveness of the proposed method.