Abstract: Forest succession simulation can reveal the restoration process and succession patterns of forest vegetation. Vegetation succession simulation in natural reserve is an important theoretical basis for forest conservation. The result of traditional simulation is usually a timeseries data, which is simulated under current ecological environment, hard for validation, and not adequate for the guidance of forest conservation. Therefore, an essential part of simulation is to know how human activities affect forest succession by considering ecological conditions as model variables, and applying the simulation model under different ecological conditions. This study introduced the metapopulation theory into the forest succession simulation by taking Mount Lushan as a case study, and selected 10 dominant species for the simulation. MATLAB was used to simulate different conditions, i.e. current habitats, habitat destruction, and habitat restoration. Combining the area ratio curve of forest succession dynamics and the responses of 10 species to different conditions, this paper summarized the protection measures for different species. The results indicated that the overall succession for all forest populations trended towards evergreen broadleaved forest, and broadleaved trees were dominant in mixed forests. The dominant species in forest community could be divided into three levels: (1) The firstlevel species have strong adaptability to harsh ecological environment, but they are easy to be replaced by species with low dominance when habitats keep stable, and we should prevent artificial introduction of exotic species from excessively occupying the habitats of original dominant species; (2) The secondlevel species are sensitive to habitat change, and serious habitat damage will cause their extinction. However, they can survive under habitat restoration; (3) The thirdlevel species have difficulty to survive even under habitat restoration, and artificial cultivation measures should be taken to protect these species.