Abstract: To solve the problems of complex die surface design process for aircraft skin, dependence on design experience  and low reuse degree of process knowledge, the process parameters were extracted based on section line method to achieve the digital expression of the aircraft skin forming process feature. Then, through adopting the nearest neighbor strategy, the influences of extreme values of the process parameters were eliminated, and the stability and reliability of similarity calculations were improved by standardizing the feature matrix based on interpolation methods, improving the traditional TOPSIS algorithm and introducing a robust normalization algorithm. Furthermore, based on the above research, a rapid retrieval system for aircraft skin forming process of process feature recognition and similarity matching was developed to achieve fast retrieval and knowledge reuse of aircraft skin process parts. The results of on-site application testing in enterprise show that the process search accuracy of system reaches 90.5%, significantly improving the efficiency and quality of process design for aircraft skin parts.