Abstract Purpose Soil microbial communities and related key ecological processes play critical roles in timberline delineation and soil carbon balance in alpine ecosystems, which are highly vulnerable to climate change. Accordingly, understanding their geographical differentiation will facilitate recognition of ecosystem functions and improve soil carbon models. In this study, we explored the biogeographic patterns of soil bacterial communities and their mechanisms in maintaining soil carbon balance in an alpine timberline ecosystem of the Sygera Mountains, Southeast Tibet. Materials and methods Soil samples were collected from typical forest belts above and below the timberline. The abundance and composition of bacterial communities, as well as functional genes, were assessed using the gene chip technology. The relationship of key microbial taxa, functional genes, and soil carbon maintenance was investigated using random forest analysis, multi-model inference, and structural equation modeling. Results and discussion The shrubland soil bacterial community exhibited greater diversity compared with the coniferous forest community, with higher Shannon Index and more functional genes at the taxonomic and functional levels, respectively. Bacterial community composition differed between the two forest types, with copiotrophic bacteria more abundant in shrubland, and oligotrophic bacteria more abundant in coniferous forest. The shrubland community was also more efficient at utilizing labile organic carbon, while the coniferous forest community utilized recalcitrant organic carbon more efficiently. Genes related to labile carbon degradation were more intense in shrubland, while genes related to recalcitrant carbon degradation were more concentrated in the coniferous forest. Soil temperature and C:N ratio were dominant drivers of bacterial community composition and function. Besides key soil-environment and microbial properties, certain bacterial taxa and functional genes also exerted unique roles in soil carbon variation. Conclusions Significant differences exist in soil bacterial community composition and functions between the two forest types above and below the timberline of the Sygera Mountains. These differences may be attributed to soil temperature and soil C:N ratio. Coupling these microbial variables into the earth system model can improve the predictive power of the carbon feedback process in terrestrial ecosystems.