The gut microbiome in humans is associated with geography, diet, lifestyles and so on, but its relationship with some isolated populations is not clear. We used the 16sRNA technique to sequence the fecal microbiome in the Chinese isolated Yao population and compared it with the major minority Zhuang and the major ethnic Han populations living in the same rural area. Information about diet frequency and health status and routine serum measurements were collected. The unweighted UniFrac principal coordinates analysis showed significant structural differences in fecal microbiota among the three ethnic groups. Statistically significant differences were observed in the community richness estimator (chaos) and the diversity estimator (Shannon) among the three groups. At the genus level, the fecal samples of the isolated Yao population presented the lowest relative abundance of the Megamonas genus, which was potentially related to the high frequency of bean consumption in the diet. Two enterotypes were identified in the overall fecal microbiota in the three populations. In the isolated Yao population, a higher Bacteroides abundance was observed, but the Prevotella abundance decreased with increased alcohol consumption.