Worldwide, private lands have attracted increased attention from conservationists, not only because most of the globe is privately owned, but also because private lands can be an asset to the protected area conservation strategy. In Brazil, the riverine Areas of Permanent Protection (APPs) is a key instrument of the Forest Code to protect native vegetation on private lands. Although APPs were conceived to function as potential wildlife corridors, this putative role has been rarely assessed. Further, recent debatable changes in the Forest Code have decreased the extent of APPs. Given this lack of information and the declining area being protected along with the growing demand for agricultural and biofuel production, which stimulate additional deforestation, it is timely to assess the role of APPs in wildlife conservation. We did this analyzing how ocelot (Leopardus pardalis) occurrence is influenced by covariates in 3 cerrado landscapes dominated by sugarcane and managed forests of eucalyptus and pine trees. We collected detection/non-detection data by camera trapping during 2 dry seasons (2013 and 2014) at 208 stations (6606 camera-days). We estimated ocelot detection and mean relative abundance using a single-species/single-season occupancy model that accounts for heterogeneous detection probability induced by variation in abundance. Modeling results showed that percentage of native forest was the most important covariate to explain ocelot mean relative abundance. This parameter was also affected positively by APPs, with ocelots being more abundant inside than outside APPs and, unexpectedly, by sugarcane. Given study design, however, the positive effect of this crop likely reflects the contact zone between sugarcane and native forest. Our findings show that landscape composition affects ocelot abundance and highlight the importance of APPs and Legal Reserves in agricultural landscapes. We conclude that, in such landscapes, compliance to the Forest Code by private land owners is key to supporting ocelot occurrence.