Abstract: In order to accurately extract the characteristic parameters of the welding seams in the digital model of ship welding parts, and then to realize the adaptive and rapid matching and selection of welding processes in the robot database system, a feature parameter extraction algorithm of ship welds based on spatial position and contour distance is proposed. First, the spatial position relationship of the welded joint is determined by the recognition of the surface to be determined based on the Helen formula, and the weld feature is recognized through the combination of the minimum contour distance; then the feature points and lines at the weld groove are identified based on the total number of contour lines and the two end points with the smallest outline distance; finally, the final weld feature parameters related to the welding process are extracted based on the mathematical model built through the three types of welded joints. The test results show that the proposed welding seam feature parameter extraction algorithm can accurately identify 4 types of joint forms and 10 groove types, as well as accurately extract parameters such as weld gap, groove included angle and welded plate thickness, which is characterized by wide weld feature recognition and complete information extraction. Compared with other related recognition algorithms, the recognition rate of the proposed algorithm reaches 100%, and the recognition efficiency is increased by 16.06%, which further verifies the effectiveness of the algorithm.