Abstract: The assembly accuracy of aircraft buckle parts directly affects the shape accuracy and airtightness of the fuselage skin. To address the problem of difficulty in accurately extracting key features during digital repair and measurement of buckle parts, a method for extracting key corner features of aircraft buckle parts contour based on 3D point cloud was proposed. Firstly, the obtained 3D cloud data was preprocessed to obtain 2D point cloud data after dimensionality reduction. Secondly, the precise extraction of the assembly contour of the aircraft buckle parts' buckle plates and complementary slots was achieved by using the latitude and longitude method and the "convex hull removal+density gradient screening" method. Then, the dual nearest neighbor search method was used to obtain the candidate corner features of the assembly contour. Finally, based on the euclidean clustering method and hash mapping method, the extraction of key corner features of buckle plates and complementary slots was achieved. The experimental results show that the method proposed has better extraction performance and operational efficiency compared to existing methods, and has good robustness, which can effectively extract key corner features of aircraft buckle parts contour.