The reconfigurable test platform (RTP) is designed to test high speed, high reliability and super precision indexes of wafer transmission robot, which can transfer and align wafers in the semiconductor manufacturing. This paper focuses on optimal configuration of the test platform. First, the problem is described considering three important yet conflicting factors: assembly, structural stiffness, and cost. Second, based on the combination sequence of function requirements and the connection relationship of modules, a multi-objective optimal decision-making model was formulated. And then, based on layered network and dependent degree methods, the multi-objective model is improved to be an only one objective optimization problem. Finally, the modified discrete particle swarm optimization (MDPSO) algorithm is proposed to solve the model and generate the best configuration scheme of RTP. Our computational results have shown that the MDPSO algorithm is efficient and robust. And the results demonstrate that the developed methods can be used to improve the test platform design efficiency and reduce the production cost.