Soil attributes may affect the residual activity and impacts of herbicides on non-target organisms. This study aimed to evaluate the residual activity of the sulfentrazone herbicide and its impacts on microbiological parameters of Brazilian Savanna soils, as well as to identify soil attributes that can be used to predict the residual activity of the herbicide in these soils. The experimental design was completely randomized, in a 6 x 5 factorial scheme, with 4 replications, being 5 soil types (Rhodic Acrustox; Typic Quartzpsamment; Rhodic Haplustox; Typic Humaquepts; Fluventic Dystrustepts) and 6 periods (0, 15, 30, 45, 60 and 90 days after herbicide application - DAA). The residual activity of the sulfentrazone herbicide is dependent of the soil type, and the time that the herbicide can control at least 80 % of Urochloa decumbens plants ranges from 0 (Typic Humaquepts) to 38 days (Rhodic Haplustox). The bases saturation is the only soil attribute correlated with the herbicide efficacy over time, in the Brazilian Savanna soils, showing a moderate correlation up to 90 DAA. The pH, organic matter and clay content are not correlated with the herbicide efficacy over time in these soils. Sulfentrazone reduces the microorganisms activity from 45 to 90 DAA in all the studied soils. However, it does not affect the microbial biomass carbon.