The invention provides a point cloud data automatic classification method for a power transmission line based on a random forest model. With the development and application of the technology of laser radars, an airborne laser radar is gradually introduced to the routing inspection operation of the power transmission line. The laser radar will generate a large amount of point cloud data after the scanning of a power transmission line corridor, and the classification of the point cloud data is the foundation for the subsequent analysis and processing of the point cloud data. However, a conventional manual classification method is larger in workload, and consumes a lot of time. According to the invention, the method achieves the features of machine learning based on the local features of points, and achieves the gradual calculation of all feature points, and then a learning machine is built through a random forest. The point cloud data of manual classification is taken as the training sample for the building a supervised learning model. After the model training, the model is used for the automatic classification of the point cloud data which is not classified.