This study aimed to characterise soil microbial community structure and function in temperate upland grassland ecosystems. We compared the use of community level physiological profiles (CLPP), phospholipid fatty acid (PLFA) profiles and community DNA (%G+C base distribution) approaches to quantify soil microbial community structure and potential activity across a gradient of three upland grassland types at 10 geographically distinct sites within the UK. Soil microbial biomass (Cmic) was highest in unimproved (U4a) and lowest in improved (MG6) grasslands. In contrast, potential soil microbial activity (carbon utilisation) was greatest in the improved and lowest in the semi-improved (U4b) grasslands. PLFA and culturing revealed that the soil microbial community shifted from one favouring fungi to one favouring bacteria as grassland improvement increased. Canonical variate analysis (CVA) of the CLPP and PLFA data differentiated microbial communities from the grassland types and sites and the separation between grasslands was greater using PLFA than CLPP. Discrimination between grasslands was mainly due to the presence of higher concentrations of fatty acids typical for Gram −ve bacteria in improved grasslands and actinomycete and fungal fatty acids in the semi and unimproved grasslands. CVA of the %G+C data gave less discrimination of the microbial communities than the other two methods. Correlation analysis of the CVA data for each microbial analysis showed a small, but significant, level of matching between the CLPP and PLFA data suggesting these two analyses may be reporting on similar members of the microbial community. Correlation between microbial community structure and soil physio-chemical properties indicated that PLFA were highly correlated with calcium, phosphorus, sodium, nitrogen and organic matter content and pH. CLPP were highly correlated with sodium and organic matter content and pH, while %G+C content correlated with pH. Correlation between microbial community structure and plant community structure indicated that fatty acids typical for Gram −ve bacteria were highly correlated with the presence of Lolium perenne and Trifolium repens and all microbial PLFA with the presence of Vaccinium myrtillus. Correlation of plant species with CLPP indicated that the presence of a number of rushes, shrubs, herbs and grasses influenced the metabolic profiles of the microbial communities from these grasslands. The presence of herbs were found to be highly correlated with certain %G+C classes within the community DNA.