Abstract: To detect the ecological process of the succession series of Phyllostachys glauca forest in a limestone mountain, five niche models, i.e., broken stick model (BSM), niche preemption model (NPM), dominance preemption model (DPM), random assortment model (RAM) and overlapping niche model (ONM) were employed to describe the speciesabundance distribution patterns (SDPs) of 15 samples. x² test and Akaike information criterion (AIC) were used to test the fitting effects of the five models. The results showed that the optimal SDP models for P. glauca forest, bamboobroadleaved mixed forest and broadleaved forest were DPM (x²=35.86, AIC=-69.77), NPM (x²=1.60, AIC=-94.68) and NPM (x²=0.35, AIC=-364.61), respectively. BSM also well fitted the SDP of bamboo-broadleaved mixed forest and broad-leaved forest, while it was unsuitable to describe the SDP of P. glauca forest. The fittings of RAM and ONM in the three forest types were all rejected by the x² test and AIC. With the development of community succession from P. glauca forest to broadleaved forest, the species richness and evenness increased, and the optimal SDP model changed from DPM to NPM. It was inferred that the change of ecological process from habitat filtration to interspecific competition was the main driving force of the forest succession. The results also indicated that the application of multiple SDP models and test methods would be beneficial to select the best model and deeply understand the ecological process of community succession.