Building change detection provides powerful support for building management and decision making. The current challenge of building change detection is the high calculated cost of completing the task. In order to improve the detection efficiency, a method based on Potree structure for change detection between the laser point cloud and BIM point cloud of a building was proposed. The method compare the laser point cloud acquired in real-time with the BIM planned at the start of the construction, detecting and identifying the differences between them as a result of building change detection. The experimental results show that the proposed method in this paper reduces the time complexity by 22.05% compared to the modifiable nested octree based method, while ensuring no loss of completeness and accuracy.