A change in precipitation can profoundly change the structure of soil microbial communities, especially in arid and semi-arid areas which are limited by moisture conditions. Therefore, it is crucial to explore how soil bacterial community composition and diversity will respond to variation in precipitation. Here we conducted a precipitation control experiment to simulate precipitation change by reducing and increasing rainfall by 25%, 50%, and 75% in the alpine grasslands of northern Tibet. The composition, diversity, and species interaction network of soil microbial community were studied by high-throughput sequencing, and the relationship between microbial community species and soil environmental factors was analyzed. Our results showed that Proteobacteria (45%-52%) and Actinobacteria (37%-45%) were the dominant bacteria in the soil. The alpha diversity index based on Shannon, Chao1, and Simpson indices revealed that precipitation change had no significant effect on richness and evenness of soil microbial communities. Non-metric multidimensional scaling (NMDS) and analysis of similarities (ANOSIM) showed that a clear separation of soil microbial communities between D2(-50%),D3(-75%) and W2(+50%), W3(+75%) treatments. The microbial interaction network indicated that the water-increasing treatment group had closer connections, and Proteobacteria and Actinomycetes were the core species. Furthermore, there was a stronger positive correlation between species in the water-reducing treatment group, the contribution of Proteobacteria decreased significantly, the role of connecting hub decreased, and Actinomycetes became the most important core microbial species. In addition, soil water content (SWC) and available phosphorus (AP) were closely related to the variations in soil microbial compositions. The findings of this study provide a theoretical basis for the driving mechanism of global climate change on soil microbial community and grassland ecosystem in alpine grassland.