Abstract: In the development of CAPP system integration, instrumentalization and intelligence, decision reasoning and intelligent technology of craft scheme are the key points. Traditional craft decision making method has the problem that it can not make intelligent choice in multi-craft scheme. To solve this problem, a craft decision reasoning method based on knowledge graph representation learning was proposed. In this method, the translation distance model is used to represent the craft knowledge vectorially, and the distance between the ideal result and the decision result is obtained by vector operation. After model training, the distance is constantly reduced, which leads to the stability of the loss function. Finally, the optimal result is obtained, and the rationality of the decision scheme is strengthened by fuzzy function calculation and analysis. The problem of intelligent decision in multi-craft scheme is solved. In this paper, P0 class 6203 bearing is used as an example to verify the proposed method. The results show that the proposed method can effectively realize the decision-making of the parts craft scheme and improve the ability of intelligent decision-making of the CAPP system.