Abstract: Accurate and rapid segmentation of individual teeth from cone beam computed tomography (CBCT) images is of great significance in dental treatment. In order to solve the problems of difficult processing of occlusal surface and low resolution of CBCT image, a multi-view fusion-based algorithm for tooth instance segmentation in CBCT images is proposed in this paper. First, the CBCT data is transformed into sequential slices in both the axial and sagittal views. Second, by leveraging the characteristics of data from different views, the segmentation results from the axial view provide guidance and instance label information for the sagittal view segmentation. Finally, the instance segmentation results from the sagittal view images are used to reconstruct the CBCT data, yielding high-precision tooth instance segmentation results. Based on Test29 structure of CBCT dataset for example, with average Dice coefficient as evaluation indicators, the experimental results show that the average Dice coefficients of teeth and occlusal surfaces are 89.74% and 93.56%, respectively, which representing an improvement of 0.82 to 3.45 percentage points compared to algorithms such as CGDNet; additionally, the average segmentation speed is 9.2 s, which realizes the task of high-precision instance segmentation of CBCT data in a short time.