In order to improve the accuracy and efficiency of power line extraction and reconstruction based on unmanned aerial vehicle (UAV)-borne light detection and ranging (LiDAR) point cloud, a comprehensive power line point cloud extraction and reconstruction method was adopted, in which a step-by-step extraction of power line point clouds, segmented k-means clustering sampling, and a combination of straight line and parabola fitting were combined together. Based on the preprocessing and filtering of the power line corridor point cloud, the power line corridor was segmented and roughly extracted according to the distribution characteristics of the power line point cloud in the elevation direction. Then, the roughly extracted power line points were projected onto the horizontal plane and extracted finely by Hough transform. Subsequently, the single-line separation of the power line was completed by performing segmented k-means cluster sampling on the extracted power line points. Finally, a model combining a straight line and a parabola was used to fit and reconstruct the power line. The measured data was selected to test, and the algorithm was evaluated by the integrity rate of power line extraction and the accuracy and efficiency of reconstruction. The experimental results show that the comprehensive integrity rate of extracting power line points using this algorithm is above 96%, and the result is helpful for efficiently extracting and reconstructing single file single power line point cloud.