Abstract: To protect the critical product design information in collaborative design,a matrix-based design problem decomposition method was proposed.For solving the problems of the dispersed distribution of protected parameters and the narrow range of choice of shared parameters in the existing function-parameter matrix clustering algorithm,a constrained clustering algorithm based on the artificial bee algorithm was proposed and an initial module partition scheme was obtained.In order to further reduce the risk of information leakage,an approach to adjusting the initial module partition scheme was proposed.A measure called the degree of relative participation was used to identify the protected functions which may lead to information leakage of the protected parameters.When such function existed,the assignment of the related functions and parameters was adjusted,to enhance the degree of relative participation.Finally,the relief valve system was used to verify the feasibility and effectiveness of the proposed method.