Abstract:Aims We tested the neutral theory of biodiversity on subalpine meadows of the eastern Tibetan Plateau that exhibited comparatively complicated species composition. Our objective was to explain the species abundance distribution pattern and the underlying mechanism of biodiversity. Methods We fit the neutral model to a randomly sampled data set obtained in three different habitats (north-facing slope, level field and south-facing slope) and used three methods to test the fitness of the neutral model to the real community: confidence interval, goodness of fit and diversity index. Important findings We found no significant difference (p>0.05) between the neutral theory predictions and observed species abundance distributions in the three habitats according to the goodness of fit method. The observed data nearly completely fall into 95% confidence intervals of the neutral model predictions (only one out of 63 species in level field communities and 2 out of 75 species in the north-facing slope communities deviate from the 95% confidence interval). There is no significant difference between the neutral theory predictions and observed species abundance patterns, in which the fit of richness predictions is the best (0.49<p<0.56) and the fitness of evenness predictions is relatively poor. However, for the three different habitats, the fitness of these three indices in north-facing slope communities is perfect and the p-values vary between 0.49 and 0.70, but the fitness in level field communities is poorer (p-value of the Simpson diversity index is less than 0.1). Although the test results of the neutral theory by three different test methods and habitats are somewhat different, we conclude that the neutral model can predict species abundance distribution patterns in the three habitats of subalpine meadow.