Abstract: A method is presented to integrate partly overlapping free from surface measured with structured light 3D vision sensor from multiple viewpoints into a whole surface model of object. Firstly, based on the given initial coordinate rotation matrix R and transformation vector T of different viewpoints, an iterative registration method is applied to calculate more accurate motion parameter R and T so that 3D data obtained from different viewpoints can be translated into a unique frame of reference. Then, by introducing spherical coordinate of 3D space , the registered data is divided into grids according to their space angle and a mesh overlaying the whole surface of object is obtained . Finally an experiment of registration and modeling is presented.