Of the threats facing sloth bear (Melursus ursinus) populations, habitat fragmentation is the most pressing. Although conservation requires protection of habitat, little is known about the factors governing sloth bear occurrence. We used camera-trapping data to investigate occupancy of sloth bears in Mudumalai Tiger Reserve, an important conservation site in India during January–April 2010. Presence–absence data, collected under a systematic sampling framework, were used to test a priori hypotheses incorporating covariates believed to influence sloth bear occurrence. The estimated occupancy of sloth bears in the study area was 0.83 (SE  =  0.01) with a detection probability of 0.23 (SE  =  0.07). We found that no model with covariates was as strongly supported as the null model, suggesting that covariates we chose were relatively weak predictors of use. That said, our results suggested that sloth bear use was associated with deciduous forests; weaker evidence was found for association with termite mounds and fruiting trees. In the future, monitoring programs for wide-ranging species could benefit from using occupancy surveys.