The structural parameters of urban landscape trees are important information for studying their phenotypes, ecological functions and landscape functions. The measurement cost is high to obtain various structural parameters of a standing tree. By establishing the relationship between tree parameters, it is possible for one parameter to predict another, which can reduce costs. In this study, the portable handheld laser radar scanning system was utilized to obtain high-precision three-dimensional point cloud data of the tree structure in Jinshan Park, Cangshan District, Fuzhou City, with the Ceiba speciosa as the research object.The target object was segmented after the point cloud pre-processing.Then, tree height, diameter at breast height, crown width and height under branches were extracted by human-computer interaction. The Pearson analysis result showed that breast height diameter was significantly correlated with tree height and crown width. Finally, the linear, logarithmic, exponential, power, binomial and trinomial function were used to fit the relationship model between diameter at breast height and tree height, crown width, respectively. According to the sample statistics and consistency test value of the fitting function, the most suitable function model was selected andits generalization ability was evaluated. The results show that the optimum model of tree height-diameter at breast height is H=8.665 36e^{0.013 13D}and width-diameter at breast height is C_W=-38.650 76+4.358 06D-0.127 89D²+0.001 24D³.