Laser scanning is one of the modern and actively-developing remote sensing techniques, resulting in a point cloud, containing a set of different attributes for each point. One of the positive features of laser scanning is the high accuracy of the results; this is achieved by obtaining a large number of points describing the scanned object. In some circumstances, point clouds may contain billions of points, which require hundreds of gigabytes to be stored. Loading and processing of such huge data require large time and computational resources. The first problem is such massive point clouds initial downloading and pre-processing. The standard approach is the sequential processing of laser scanning results, which requires a significant amount of time. In this paper, we have conducted research and testing of various approaches for loading and processing of point clouds, one of the proposed approaches is the use of multithreading to significantly reduce time. The guideline for improvement of processing of laser scanning point clouds with use of multithreading is presented.