The effect of grazing by large herbivores on the microbial community and the ecosystem functions they provide are relatively unknown in grassland systems. In this study, the impact of grazing upon the size, composition and activity of the soil microbial community was measured in field experiments in two coastal ecosystems: one salt marsh and one sand dune grassland. Bacterial, fungal and total microbial biomass were not systematically affected by grazing across ecosystems, although, within an ecosystem, differences could be detected. Fungal-to-bacterial ratio did not differ with grazing for either habitat. Redundancy analysis showed that soil moisture, bulk density and root biomass significantly explained the composition of phospholipid fatty acid (PLFA) markers, dominated by the distinction between the two grassland habitats, but where the grazing effect could also be resolved. PLFA markers for Gram-positive bacteria were more proportionally abundant in un-grazed, and markers for Gram-negative bacteria in grazed grasslands. Bacterial growth rate (leucine incorporation) was highest in un-grazed salt marsh but did not vary with grazing intensity in the sand dune grassland. We conclude that grazing consistently affects the composition of the soil microbial community in semi-natural grasslands but that its influence is small (7 % of the total variation in PLFA composition), compared with differences between grassland types (89 %). The relatively small effect of grazing translated to small effects on measurements of soil microbial functions, including N and C mineralisation. This study is an early step toward assessing consequences of land-use change for global nutrient cycles driven by the microbial community.