ABSTRACT: Spatial patterns and drivers of soil microbial communities have not yet been well documented. Here, we used geostatistical modeling and Illumina sequencing of 16S rRNA genes to explore how the main microbial taxa at the phyla level are spatially distributed in a 25-ha karst broadleaf forest in southwest China. Proteobacteria, dominated by Alpha- and Deltaproteobacteria, was the most abundant phylum (34.51%) in the karst forest soils. Other dominating phyla were Actinobacteria (30.73%), and Acidobacteria (12.24%). Soil microbial taxa showed spatial dependence with an autocorrelation range of 44.4-883.0 m, most of them within the scope of the study plots (500 m). An increasing trend was observed for Alphaproteobacteria, Deltaproteobacteria, and Chloroflexi from north to south in the study area, but an opposite trend for Actinobacteria, Acidobacteira, and Firmicutes was observed. Thaumarchaeota, Bacteroidetes, Gemmatimonadetes, and Verrucomicrobia had patchy patterns, Nitrospirae had a unimodal pattern, and Latescibacteria had an intermittent pattern with low and high value strips. Location, soil total phosphorus, elevation, and plant density were significantly correlated with main soil bacterial taxa in the karst forest. Moreover, the total variation in soil microbial communities better explained by spatial factors than environmental variables. Furthermore, a large part of variation (76.8%) was unexplained in the study. Therefore, our results suggested that dispersal limitation was the primary driver of spatial pattern of soil microbial taxa in broadleaved forest in karst areas, and other environmental variables (i.e., soil porosity and temperature) should be taken into consideration.