Based on a mobile 3D lidar, a method was established to acquire citrus canopy structure information including the tree height, the crown width, and branch angle. Firstly, using mobile 3D lidar to obtain 3D point cloud data of citrus. Secondly, select the SOR filter combination (k, α) as (100, 0.9), (20, 1.2), (100, 0.9) to filter the point cloud of the canopy incomplete leaves and the point cloud of the leaf edge. Thirdly, use the mean shift algorithm to obtain the center of the point cloud of the plant combined with the angle of the collection point to complete the initial alignment of the point cloud. Fourthly, using the ICP algorithm to accurately splicing and constructing a three-dimensional model of the complete plant, extracting three kinds of structural information such as the plant height, crown width and branch angle of citrus. The verification results show that the relative errors of tree height, crown width and branch angle obtained by the 3D model are less than or equal to 2.5%, 4.5%, and 5.5%, repectively. The determination coefficient R2 is greater than 0.97, and the root mean square error is greater than 5.1.