Habitat fragmentation can promote patches of small and isolated populations, gene flow disruption between those populations, and reduction of local and total genetic variation. As a consequence, these small populations may go extinct in the long-term. The ocelot (Leopardus pardalis), originally distributed from Texas to southern Brazil and northern Argentina, has been impacted by habitat fragmentation throughout much of its range. To test whether habitat fragmentation has already induced genetic differentiation in an area where this process has been documented for a larger felid (jaguars), we analyzed molecular variation in ocelots inhabiting two Atlantic Forest fragments, Morro do Diabo (MD) and Iguaçu Region (IR). Analyses using nine microsatellites revealed mean observed and expected heterozygosity of 0.68 and 0.70, respectively. The MD sampled population showed evidence of a genetic bottleneck under two mutational models (TPM = 0.03711 and SMM = 0.04883). Estimates of genetic structure (FST = 0.027; best fit of k = 1 with STRUCTURE) revealed no meaningful differentiation between these populations. Thus, our results indicate that the ocelot populations sampled in these fragments are still not significantly different genetically, a pattern that strongly contrasts with that previously observed in jaguars for the same comparisons. This observation is likely due to a combination of two factors: (i) larger effective population size of ocelots (relative to jaguars) in each fragment, implying a slower effect of drift-induced differentiation; and (ii) potentially some remaining permeability of the anthropogenic matrix for ocelots, as opposed to the observed lack of permeability for jaguars. The persistence of ocelot gene flow between these areas must be prioritized in long-term conservation planning on behalf of these felids.