The reconstruction of helical surface plays an important role in many engineering and scientific applications. This paper tends to establish an effective processing scheme of surface reconstruction of scatter 3D points, and make an in-depth exploration for the preprocessing technology of scatter points set and the algorithm of three-dimension model surface reconstruction. This paper has three main aspects: data acquisition, data preprocess and surface reconstruction. First, by comparing several measuring equipments, the screw parts are scanned via CMM (Coordinate Measuring Machine). Initial 3D point clouds are obtained by setting the scanning route according to the shape of the targets. Second, Using "3σ" criteria point de-noising is applied to the initial data points. Then Heap Sort is used to sort these points, being convenient for saving data and reconstructing the surface. Third, this paper presents a surface reconstruction method based on triangulation algorithm. The triangular meshes are generated based on Delaunay triangulation technique in two dimensions. And Loop Subdivision is adopted in order to get manifold meshes. Finally the surface reconstruction of an involute worm and an Archimedes worm shows that this approach is feasible.© (2010) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.