Journal of Data Processing Fu Yan, Zhai Jin Lei 6 4 2016 http://www.dline.info/jdp/fulltext/v6n4/jdpv6n4_1.pdf The Power Crust algorithm is widely used in the 3D reconstruction. But when we directly use it in the process of reconstruction, there are some problems, such as the large calculating quantity and poor denoising ability, etc. This paper puts forward a new method to solve the problems of the Power Crust algorithm. In stage I, we use the KD tree algorithm to get every point’s neighborhood information within the points cloud. In stage II, we use the Laplace method to denoise the points sets. In stage III, we use the neighborhood average method to simplify the points sets when the points are too many. At last, we use the Power Crust algorithm to accomplish the reconstruction. The result is a 3D model constructed from the simplifying points sets. In addition, compared with the previous method, this paper’s method systematically solves the denoising problem and simplifies the points cloud. It has a fast reconstruction speed and a good reconstruction model.