Point cloud data are acquired using 3D scanner, such as a terrestrial laser scanner and land-based mobile mapping systems, in a surveying, mapping, structure maintenance, and environment monitoring. The latest 3D scanners perform a rapid and massive data acquisition. However, the massive data require huge processing time in data sharing, visualization and 3D modeling. We propose a methodology to improve a performance in 3D surface modeling using point cloud data projected into a panoramic space. Thus, we have clarified that our approach can improve workability in 3D modeling. Additionally, we confirmed that acquired 3D point cloud data with terrestrial laser scanner data can be classified on 2D panoramic range image using normal vectors estimated from point cloud. This classification is based on 2D image processing. However, we also confirmed that a result from proposed modeling was equivalent to conventional 3D modeling.