Abstract:The soil microbial community, which indicates soil quality changes, has been regarded as an early warning and sensitive indicator of soil ecosystem change. The phospholipid fattyacid (PLFA) method was used to study the changing rule and response of the soil microbial community structure to different intensities of grazing and fencing of alpine rhododendron shrubland. The results showed that soil total PLFA, bacterial biomass, actinomycetes biomass, and fungi biomass were significantly reduced with an increase in grazing intensity (P < 0.05). At heavy grazing intensity, soil total PLFA, bacterial biomass,actinomycetes biomass, G+/G- ratio, and the stress index after the fenced treatment were significantly higher than those after the grazing treatment, whereas the bacteria/fungi ratio after the fenced treatment was significantly lower than that after the grazing treatment. At medium grazing intensity, soil total PLFA, bacterial biomass, and bacteria/fungi ratio were significantly higher than those after the grazing treatment, whereas fungal biomass and the stress index were significantly lower than those after the grazing treatment. At light grazing intensity, biomass and biomass ratio after the fenced treatment and those after the grazing treatment were not significantly different. PLFA principal component analysis (PCA) showed that 53.68% of PC1 was composed of straight-chain saturated and unsaturated fatty acids 14:0, 15:0, 10Me16:0, and 18:1ω9c, whereas 51.34% of PC2 was composed of branched-chain saturated and cyclopropane fatty acids i16:0, 16:1ω7c, i17:0, cy17:0, 17:0, 18:1ω9t, 18:0, and cy19:0. Grazing grassland and fenced grassland of different grazing intensities had similar soil microorganisms. The fenced treatment affected the microbial community structure to a greater extent than did the grazing treatment. Correlation analysis showed that soil total PLFA, bacteria biomass, fungi biomass, and actinomycetes biomass were significantly and positively correlated with soil organic carbon and total nitrogen at P < 0.001, whereas bacteria/fungi had a significant and negative correlation with soil organic carbon and total nitrogen (P < 0.001). The above indicated that overgrazing reduces soil microbial activity of alpine shrublands and significantly reduces the soil microbial biomass, the soil microbial community was stable in fenced and moderately grazed areas, and fencing is advantageous to recovery of the soil microbial community in overgrazed grasslands.