In order to solve the performance optimization problem caused by the high-dimensional nonlinear characteristics of the trajactory correction decoy and improve the traditional serial design method in the conceptual design stage, an intelligent optimization algorithm based on DOE (design of experiments) and RSM (response surface methodology) was proposed and the basic projectile structure model and related design parameters were defined. Based on the DOE, the design variables were mapped to the performance criteria to generate the stochastic Kriging response surface, and the neural network was trained to identify the unstable design. Multi population genetic algorithm was used to determine the optimal projectile design. By changing its cost function, the Pareto frontier reflecting the performance tradeoff could be generated. Simulation results show that as for the technical transformation of infrared interference decoy based on trajectory correction, the proposed algorithm can obtain the optimal design configuration quickly and accurately in the conceptual design stage, which can guarantee the following accompanying flight mission.