To enhance the milling quality of aircraft skin and meet the precision requirements for butt joint gaps during assembly, we propose a method for extracting the milling auxiliary line of aircraft skin based on multi-feature selection. Initially, the collected point cloud data undergoes preprocessing through bilateral filtering. Following this, a hierarchical search structure that combines multiple point cloud features is employed to search for boundary points. The spatial tangent continuity feature is applied to extract the primary part of the point cloud boundary points. Subsequently, an optimized local surface standard deviation feature is utilized to search for the remaining points, and the results are merged to obtain the complete set of boundary points. Finally, an anisotropic optimization algorithm is applied to contract the scattered boundary points into linear shapes. Experimental results indicate that the precision of the detected boundary points can reach 0.95, with an error margin of less than 0.3 mm. When processing the skin workpiece by using the boundary points obtained by the method in this paper as auxiliary lines, the average value of the butt joint gap after splicing is less than 0.4 mm.