Quercus acutissima, a broad-leaved deciduous tree of genus Quercus (oaks) in the family Fagaceae, is one of the zonally tree species in Qiaoshan of southern the Loess Plateau. There are three main types of Q. acutissima communities in this region: pure forests of Q. acutissima (PF), mixed forests of Q. acutissima with broad leaved tree like Q. wutaishansea and Q. aliena (QBF), and mixed forests of Q. acutissima and Pinus tabulaeformis (QPF). These forests are essential for regional ecological security. However, due to unreasonable initial planning and the lack of management for decades, most of the local Q. acutissima forests have been seriously destroyed with unclear inter-and intra-relationship between species. In order to investigate the succession dynamics of the main population in Q. acutissima forests and to elucidate the stabilization mechanism of the natural secondary forests, the age structure, life expectancy and survival curve of four tree populations with the highest important values in different communities (PF, QPF and QBF) were analyzed, and the succession process in the next 20, 30, 40, and 50 years were predicted by time series prediction models based on the age structure of these population. The results showed that (1) Except for Q. aliena population in Q. acutissima and P.tabulaeformis mixed forest and Acer ginnala population in Q. acutissima pure forest, other tree populations with higher importance values had better survival ability in Q. acutissima forests of different forest types; (2) The survival curves of all dominant populations were of Deevey-Ⅲ, except for the Pyrus betulifolia population in pure forests of Q. acutissima and the Q. aliena population in the mixed forests of Q. acutissima with broad leaved tree; (3) The number of young individuals of tree populations with higher importance values in different Q. acutissima community types was relatively abundant. Time series analysis showed that in the next 20, 30, 40 and 50 years, the four tree populations with the highest importance values in the three community types showed no signs of decline. In conclusion, all three types of communities could developed sustainably, and pure forests of Q. acutissim showed a tendency to evolve into broadleaves mixed forest, which indicated that the further forest management of Q. acutissima forests should all be based on the close-to-nature forest management that follows the natural succession law. Our results might provide a theoretical basis for sustainable development and management of natural secondary of Q. acutissima in southern the Loess Plateau.