Reflection intensity information is an important component of data acquired by airborne light detection and ranging (LiDAR) system. The main difficulty in realizing road point cloud extraction using reflection intensity information is that road point cloud is always contained in ground point cloud, whereas the reflection intensity threshold to discriminate road point cloud and non-road point cloud is hard to obtain. To solve this problem, an extraction algorithm of road point cloud based on skewness balancing is proposed, which can obtain the reflection intensity threshold automatically, accurately and parameter-freely. After that pure road point cloud can be obtained. We adapt two datasets of LiDAR point cloud located in somewhere of Shanxi province and a certain city in Germany to test the effectiveness of the proposed algorithms at different environments, respectively. The experimental results show that the proposed algorithm is easy and effective, and has a good performance of road point cloud extraction.