Extracting three-dimensional model of the pylon from aerial LIght Detection And Ranging (LiDAR) point clouds automatically is one of the key techniques for digitization and visualization of smart grid facilities. This paper presents a model-driven three-dimensional pylon modeling method using airborne LiDAR data. On the basis of in-depth study of the actual structure of the pylon and the characteristics of point clouds data, a conceptual model of pylon is constructed, in which the pylon is divided into three parts as pylon foot, pylon body and pylon head. Parameters of the model such as position and orientation are defined. In this approach, a complicated pylon is divided into three relatively simple parts firstly. Then different parts of the pylon are reconstructed with different strategies. Finally, model parts are assembled to a complete pylon model using the position and direction information. Results of experiments on the point clouds data from Southern Power Grid show that the precision of extracted pylon orientation and position reached centimeter-level, the accuracy of pylon head classification is higher than 95%, and the pylon model fits well with pylon points. It suggests that the proposed approach can achieve the goal of semi-automatic three-dimensional modeling of the pylon effectively.