Abstract: Based on the lidar point cloud data and ground survey data of 200 Chinese fir trees from 6 standard plots in Yangkou National Forest Farm, Shunchang County, Fujian Province, the canopy height model (CHM) generated by airborne Lidar point cloud data was used to detect the crown apex and extract the tree height, and the watershed algorithm of labeled extreme value was used to estimate the crown area. Combining the estimated tree height and crown area with the true value of individual tree stock, a single tree stock estimation model based on Catboost algorithm is constructed. The results show that using the local maximum algorithm to estimate the tree height, R² is 0.91, RMSE is 0.81 m; The watershed algorithm with marked extremum was used to estimate the crown area, with R² of 0.81 and RMSE of 1.18 m². The R² of single wood stock estimation model constructed by Catboost algorithm is 0.934. In conclusion, airborne Lidar point cloud data can effectively estimate tree height and canopy area, and Catboost algorithm can be used to estimate individual wood stock of Chinese fir, providing a new idea for high-precision forest stock inversion.