The invention relates to a 3D data based aircraft surface manufacture quality detection method. The method comprises that a series of sine stripes are projected to a surface sheath of an aircraft; left and right cameras collect stripe images modulated by the surface; the phases of the stripe images are expanded; the left and right images are matched according to phase and epipolar geometrical constraint; a 3D point cloud model of the surface of the aircraft is reconstructed; a rivet region of interest is extracted, and according to the correspondence between an image area and the 3D model, rivets are positioned in the point cloud model; a curved surface fitting method is used to obtain a reference configuration of the surface of the aircraft; the distance between a point of the rivet to the reference configuration is calculated, and a region of interest of concave-convex rivets is obtained; and a pseudo color image is coded and projected to the surface of the aircraft, and a visual result is formed. Thus, depth information of the rivet region of interest can be detected, the robustness and precision are high, a quantified and visual detection result can be provided, and a new approach suitable for the rivet region of interest of the aircraft surface is provided.