Abstract: Abstract: Fine-scale geometric modeling of actual trees can be an essential prerequisite for forest information and phenotypic characteristics at present. It is necessary to effectively map the tree architecture, and accurately retrieve the plant growth properties from light detection and ranging (LiDAR) data in a wide range of biophysical and ecological processes. Here, a synergistic approach was proposed for the skeleton modeling and property retrieval of the rubber tree using deep learning and computer graphics. The backpack laser scanning data was also collected for three typical rubber tree clones, i.e., PR107, CATAS 7-20-59, and CATAS 8-79. Firstly, the labeled wood and leaf point samples were collected using traditional machine learning assisted by manual segmentation. The acquirement strategy was coupled with the voxelization and data augmentation to achieve a suitable and larger training data set than before. Secondly, the deep learning network of reformulation architecture was developed to comprise four features encoding and decoding layers. The improved PointConv modules were embedded to calculate the local translation-invariant point features and the interpolation modules, further propagating the features from sub-sampled point clouds to a scale-up resolution. As such, both the global and local features of the forest points were comprehensively extracted to greatly reinforce the leaf-wood point cloud classification from the forest scene geometry. Thirdly, the tree skeletons were reconstructed for the spatial connectivity and cylinder fitting scheme using the extracted branch points after the classification of forest point clouds. Finally, the affiliation relationship between foliage points and the first-order branch points was determined to simplify each tree crown with as many foliage clumps and different branch compositions. The overall accuracy of the branch and leaf classification was 90.32% using the deep learning network during the field measurements of the test samples, including three rubber tree plots, which was approximately 10% higher than before. The skeleton reconstruction results show that the average value (48.07o) of the angles between the trunk and the first-order branches for rubber tree clone PR107 was larger than that (36.31o) of clone CATAS 7-20-59. The reason was that the rubber tree clone PR107 presented a spread-out crown with a larger crown volume than the clone CATAS 7-20-59 with a vase shape tree crown. Rubber tree clone CATAS 8-79 presented the largest diameter at the breast height, but the vulnerability to chilling injury resulted in considerable defoliation, even to the smallest volume of the tree crown. Meanwhile, the high estimation accuracies of the first-order branch diameter (R2≥0.94, RMSE<3.01 cm) and the included angle between the trunk and the first-order branches (R2≥0.91, RMSE≤4.94o) were achieved for the three test rubber tree plots. In addition, there was a positive correlation between the volume of each foliage clump borne on the first-order branches and the diameter of the corresponding first-order branch. Overall, the improved 3D mapping approach can be widely expected to quantitatively characterize the structural variables of rubber trees for the stand components and the plant parameter inversion using deep learning and computer graphics. The finding can also provide a novel concept for the intelligent processing of forest point clouds.