Understanding biogeographical patterns and underlying processes of belowground community assembly is crucial for predicting soil functions and their responses to global environmental change. However, little is known about potential differences of belowground community assembly among bacteria, fungi, protists and soil animals, particularly for alpine ecosystems. Based on the combination of largescale field sampling, high-throughput marker-gene sequencing and multiple statistical analyses, we explored patterns and drivers of belowground community assembly in alpine grasslands on the Tibetan Plateau. Our results revealed that the distance-decay rates varied among trophic levels, with organisms of higher trophic level having weaker distance-decay pattern. The spatial and environmental variables explained limited variations of belowground communities. By contrast, the stochastic processes, mainly consisting of dispersal limitation and drift, played a primary role in regulating belowground community assembly. Moreover, the relative importance of stochastic processes varied among trophic levels, with the role of dispersal limitation weakening whereas that of drift enhancing in the order of bacteria, fungi, protists and soil animals. These findings advance our understanding of patterns and mechanisms driving belowground community assembly in alpine ecosystems and provide a reference basis for predicting the dynamics of ecosystem functions under changing environment.