AbstractWith the development of 3‐dimensional (3D) scanning technologies in dentistry, high accuracy optical scanning data from the crown and cone beam computed tomography data from the root can be acquired easily. In many dental fields, especially in digital orthodontics, it is useful to fuse the data from the crown and the root. However, the manual fusion method is complex and difficult. A novel automatic fusion method for 2‐source data from the crown and the root was researched, and its accuracy was evaluated in this study.An occlusal splint with several alumina ceramic spheres was fabricated using heat‐curing resin. A multipoint (center of each sphere) alignment method was performed to achieve rapid registration of the crown data from optical scanning and the root data from cone beam computed tomography. The segmentation algorithm based on heuristic search was adopted to perform extraction and segmentation of the crown from whole optical scanning data. The level set algorithm and the marching cubes (MC) algorithm were used to reconstruct digital imaging and communications in medicine data into a 3D model. A novel multisource data fusion algorithm, which is based on iterative Laplacian deformation (ILD), was researched and applied to achieve automatic fusion. Finally, the 3D errors of the method were evaluated.The 3 groups of typical tooth data were automatically fused within 2 seconds. The mean standard deviation was less than 0.02 mm.The novel method can aid the construction of a high‐quality 3D model of complete teeth to enable orthodontists to safely, reliably, and visually plan tooth alignment programs.