A wide array of microorganisms survive and thrive in extreme environments. However, we know little about the patterns of, and controls over, their large-scale ecological distribution. To this end, we have applied a bar-coded 16S rRNA pyrosequencing technology to explore the phylogenetic differentiation among 59 microbial communities from physically and geochemically diverse acid mine drainage (AMD) sites across Southeast China, revealing for the first time environmental variation as the major factor explaining community differences in these harsh environments. Our data showed that overall microbial diversity estimates, including phylogenetic diversity, phylotype richness and pairwise UniFrac distance, were largely correlated with pH conditions. Furthermore, multivariate regression tree analysis also identified solution pH as a strong predictor of relative lineage abundance. Betaproteobacteria, mostly affiliated with the 'Ferrovum' genus, were explicitly predominant in assemblages under moderate pH conditions, whereas Alphaproteobacteria, Euryarchaeota, Gammaproteobacteria and Nitrospira exhibited a strong adaptation to more acidic environments. Strikingly, such pH-dependent patterns could also be observed in a subsequent comprehensive analysis of the environmental distribution of acidophilic microorganisms based on 16S rRNA gene sequences previously retrieved from globally distributed AMD and associated environments, regardless of the long-distance isolation and the distinct substrate types. Collectively, our results suggest that microbial diversity patterns are better predicted by contemporary environmental variation rather than geographical distance in extreme AMD systems.