This study evaluated the genetic structure of wild populations of the endangered primate, Leontopithecus chrysomelas. We tested the assumption that populations of L. chrysomelas, given their larger population size and a higher degree of habitat continuity, would have higher genetic diversity and less genetic structuring than other lion tamarins. We used 11 microsatellites and 122 hair samples from different locations to assess their genetic diversity and genetic structure, and to make inferences about the isolation by distance. The overall expected heterozygosity (0.51 ± 0.03) and the average number of alleles (3.6 ± 0.2) were relatively low, as is the case in other endangered lion tamarins. Genetic clustering analyses indicated two main clusters, whereas the statistical analyses based on genotype similarities and Fst suggested further substructure. A Mantel test showed that only 34% of this genetic differentiation was explained by the linear distance. In addition to linear distance, structural differences in the landscape, physical barriers and behavioural factors may be causing significant genetic structuring. Overall, this study suggests that these populations have a relatively low genetic diversity and a relatively high population genetic structure, putting in question whether the presence of agroforest systems (known locally as cabruca) is enough to fully re-establish functional landscape connectivity.