Abstract: Taking natural broad-leaved mixed forest in Jinpen Mountain Forest Farm in Xinfeng County, Jiangxi Province as the research object, taking selective felling as the main measure, spatial structure optimization as the main goal, and non-spatial structure as the primary constraint, determine the felled trees and simulate the felling of the trees in the sample plot, and optimize the forest stand. The results of the study showed that in the initial state, the total mixed degree of the natural broad-leaved mixed forest was 0.5486, the size ratio was 0.4845, the angular scale was 0.5464, and the spatial structure index was 22.4674. The existing structure of the forest stands in competition among trees. The problem of high intensity and the distribution of trees in clusters. Based on the optimization goal and constraint conditions, 39 candidate trees were selected from 420 trees in the sample plot for simulated logging, and the logging intensity was 9.286%. Compare the space before and after logging. Structural indicators, the total mixing degree(M) of the plot has increased by 0.1641%, the size ratio(U) has decreased by 0.2477%, the angular scale value(W) has decreased by 8.4919%, and it is closer to 0.5. The overall spatial structure has been optimized. This study proposes a tending logging method based on the optimization of the spatial structure of the Jinpen Mountain Forest Farm, taking into account the spatial distribution of species, and comparing and analyzing the spatial structure indicators before and after logging, which proves the scientific rationality of the logging method. It provides a reference for the tending and management of natural broad-leaved mixed forest in Jinpen Mountain Forest Farm.