The quality of aircraft skin subsidence has great influence on the overall weight of aircraft, and plays an important role in maintaining the appearance of the surface. In the manufacture of modern aircraft skins, the accuracy of the aircraft skin subsidence is also raised. In this paper, a method of extracting the feature points of the skin subsidence based on T-Scan measurement is proposed in order to achieve the purpose of digital detection of the plane subsidence characteristics of the aircraft. Firstly, scanning lines are identificated from point clouds, and the 3D point cloud data is reduced to two dimensional data. Subsequently, the first and last derivatives of each point are computed by first order difference, the sag points are extracted by judging the vector angle, and the the boundary points are found by comparing the curvature of the sag points; Finally, the boundary points are projected to the line, which is fitted with data points on upper surface, projection points will be obtained as feature points of skin sag. The method directly deals with the scan line cloud data obtained by T-Scan scanning; Through the measurement and processing of actual aircraft skin, the experiment shows that the method has good practicability, the extraction precision of sag feature points is higher.