Alpine grasslands, an important terrestrial ecosystem, play an essential role in global carbon and nitrogen cycles. Recently, overgrazing in alpine grasslands, particularly on the Tibetan Plateau, a sensitive district subject to climate change, is a serious issue leading to their degradation and degeneration, Fencing to exclude grazers has become an important management practice to protect and re-establish the degraded grasslands on the TP. Previous studies on the alpine grassland ecosystem have focused on a certain aspect. However, a systematic study has largely been lacking. The current research was focused on the severely degraded alpine grassland in Gulu Town, Naqu County, located in the north of TP. In order to determine the effect of restoration, we studied the short-term (three-year) effects of fencing on the vegetation characteristics, soil physicochemical properties, soil enzyme activities, soil microbial biomass, and soil microbial communities, and compared them with areas where grazing was allowed. Soil enzyme activity was detected using an UV spectrophotometer. Microbial biomass was measured by phospholipid-derived fatty-acid analysis (PLFA) method. Furthermore, we also studied profiling of complex soil bacterial communities based on 16S ribosomal RNA gene (rDNA) sequences by means of Next Generation Sequencing (NGS) technologies at the Ion Personal Genome Machine (PGM) platform. After fencing for three years, the alpine grassland ecosystem differed noticeably between the fencing and grazing areas across the study sites. Our study showed that: (1) The total plant cover, average vegetation height, and aboveground biomass of vegetation determined during the above ground vegetation surveys were significantly higher in fenced areas than the areas where grazing was permitted (P<0.01). However, the diversity of vegetation (Shannon-Weiner index) in the fenced area was significantly lower than that in grazing sites (P<0.01). (2) Soil pH perceptibly decreased while the dissolved organic carbon in soil was increased after fencing. Soil structure, in terms of the ratio of sand to silt was also affected with fencing (P<0.05). (3) Soil enzyme activities were not influenced by fencing. (4) Soil microbial biomass (Bacteria, Actinomycetes, and Fungi) significantly increased in the fencing area(P<0.05). (5) There was an increasing tendency in the diversity of bacteria in soil, and the community composition of bacteria was changed at the phylum level, but there was no significant difference after fencing. (6) Further, the Mantel's test suggested that the major environmental factor,soil organic matter such as total organic carbon (TOC), total nitrogen (TN), carbon to phosphorus ratio(C/P), and nitrogen to phosphorus ratio(N/P)were affecting microbial community in the fenced sites. From these results, it can be inferred that, in the grasslands ecosystem, soil sub-system was steadier than plant community sub-system. In soil sub-system, soil microbial communities were less susceptible than soil profile characteristics in fenced grasslands. In conclusions, fencing was a valid method in reconstructing aboveground biomass in the alpine grassland on TP, thereby inducing a number of changes in the soil microbial community, including the composition and diversity of microorganisms. In future, more destroyed grassland should be fenced in order to preserve the balance of the ecosystem and to resist the effects of global climate change. More studies were needed to be conducted over longer periods of fencing to further our understanding of the effects on fencing on restoration of alpine grasslands.