Ocelots (Leopardus pardalis) and other wild cats are often surveyed using camera traps to identify individuals for density estimation via capture-recapture analyses or estimate occupancy via detection/non-detection analyses. Though attractants are sometimes used in such surveys, there have not been any evaluations of the effectiveness of common visual and olfactory attractants in field settings. As part of a medium and large mammal camera survey in the San Juan – La Selva Biological Corridor, Costa Rica, we integrated camera trap data within an occupancy modelling framework to estimate the effects of hanging compact disks (visual), cologne (olfactory), and sardines in oil (olfactory) on ocelot detection probabilities. Compact disks appeared to have the most information-theoretic model support, whereas cologne received less model support. The use of compact disks in surveys was also less time-consuming and less expensive than the olfactory attractants. Ocelots are visual hunters and using visual attractants can increase detection probabilities and therefore reduce uncertainty and/or reduce survey effort to obtain robust population or occupancy estimates, although using cologne might also have similar effects. Depending on logistic constraints, we recommend employing several attractants as the most appropriate way to survey ocelots and other rare felids in the future when detection biases are assumed to be strong, particularly as part of mammal community surveys.