In order to improve the processing speed and display efficiency of laser scanning data, octree structure was constructed by parallel threshold segmentation and massive point clouds were processed in blocks. The data in octree leaf nodes was simplified by data thinning method based on height difference layer by layer and was stored in octree external memory structure. The multi-resolution level of detail model of point cloud was constructed. The internal and external memory scheduling strategy based on view change and paging database was adopted. The laser scanning point cloud data acquired by a group of power line patrol helicopters were experimentally validated. The results show that this method has advantages of octree construction speed and display efficiency of large amount of point cloud data. It can satisfy the timeliness requirement of power line patrol very well.