Abstract: In order to solve the problems of single change path and extension difficulty in existing module change methods, an adaptive module modeling method is proposed. The graph theory is applied to develop an adjacency matrix(design structure matrix) to describe the relevance between different parameters. By analyzing clustered design structure matrix, module's potential changing paths will be forecasted. The paper organized changing paths and related information into structured storage files to improve the extendibility of the changing paths and their related information. Then the structured information and fully constrained module are integrated to develop the self-adaption module. For the different changing sources, different self-regulation algorithms are proposed for the self-adaption module. Finally, the study verifies the feasibility and rationality of the proposed method with an application case of truck bumper.