Abstract: This study constructed a stand mean height inversion model based on LiDAR data, aiming to provide dynamic monitoring of the growth of Sonneratia apetala. Taking the mangrove wetland of Sonneratia apetala in the Maowei Sea of Beibu Gulf as the study object and based on the height and intensity parameters extracted using airborne LiDAR data, this study compared three models, namely random forest, support vector machine, and neural network, based on the coefficient of determination (R2), root mean square error (RMSE), Akachi information criterion (AIC), and Bayesian information criterion (BIC) and then estimated the mean height and spatial distribution of Mangrove in the study area using the optimal model. The results are as follows. The stand mean height of Sonneratia apetala in the study area is 3.90~11.58 m, and Sonneratia apetala with a higher tree height and a larger diameter at breast height is mainly distributed near the tidal trench and the middle part of the study area. In the estimation of the stand mean height of Sonneratia apetala, the maximum percentile height (hmax) had the highest contribution rate, followed by 75%~99% percentile height. The random forest regression model yielded the highest precision (R2=0.938 1,RMSE=0.58 m,AIC=80.50, and BIC=49.05), followed by the support vector machine model (R2 = 0.766 5 and RMSE = 1.27 m in the test stage), and the neural network regression model yielded the worst fitting effects. Overall, the random forest model is the optimal model for the inversion of the stand mean height of Sonneratia apetala in the study area.