The aim of this study was to explore the soil microbial variability within different forest ecosystems (evergreen broad-leaf forest (EBF), coniferous forest (CF), subalpine dwarf forest (SDF) and alpine meadow (AM) at different altitudes in mid-subtropics of China. The phospholipid fatty acid (PLFA) method was used to analyze the microbial communities in rhizosphere soil under different forest types. The relationships were also analyzed between the microbial diversity and soil nutrients. A total of 27 PLFA biomarkers were detected and the PLFA concentrations decreased in the sequence of bacteria > fungus > actinomycete > protozoa in all forest types. The microbial communities in the soil under all forest types were distinct. The predominant microflora in all soils were 18:1ω9c, 16:1ω7c, cy19:0, a17:0 and 18:0. The indexes of Simpson, Shannon-Wiener and Brillouin of soil microbial community diversity in these four forest types all showed a trend of EBF > CF > SDF > AM. According to principal component analyses (PCA), the variable variances of principal components 1 and 2, which were related to the PLFA biomarkers of soil microorganisms, were 67.67% and 17.91%, respectively. Furthermore, the total PLFAs of different soil microbial groups showed a correlation with soil nutrients and enzyme activities in all forest types. The soil microbial diversity gradually decreased in the order of EBF > CF > SDF > AM in the Daiyun Mountains. Different vegetation types affect soil microbial community composition and diversity by changing the soil physicochemical properties and enzyme activity.