The traditional measurement method of vegetation indicators consumes a lot of manpower, material resources and time, and ignores the ecological benefits of different vegetation types and structures. The point cloud data can effectively evaluate the 3D vegetation parameters of urban green space, but due to the influence of data accuracy, algorithm software and other factors, the accuracy of the results of this method has an error of 10% to 20%. In this paper, taking Tangfeng South Road in Tangshan City as an example, three typical sample plots (50 m×50 m) were selected to collect UAV remote sensing images and handheld lidar data, and a three-dimensional vegetation parameter system was constructed from the single tree scale and the sample plot scale. Correction coefficients were proposed for eight parameters in the UAV point cloud data, including the number of individual trees, the total crown width area, the green amount of the tree layer, the green amount of the shrub layer, the average three-dimensional green density, the canopy density, the gap rate and the leaf area, and the five parameters of the handheld lidar data, including the average tree height, the average crown diameter, the average crown width area, the green content of the arbor layer and the average three-dimensional green density, in order to improve the accuracy of the results and provide help for the follow-up research. Finally, this paper discusses the possible application directions of UAV photogrammetry technology and lidar technology in the future.