Kin structure, the spatial aggregation of related individuals, impacts many processes important to conservation (e.g. inbreeding), and patterns of kin structure could be impacted by human-mediated habitat fragmentation and loss. While kin structure is expected to increase with habitat fragmentation (reduced connectivity), the effects of habitat loss (reduced resource availability) remain unclear. Disentangling the effects of habitat fragmentation and loss is challenging because they usually are spatially correlated, and because most species are negatively affected by both processes. Raccoons Procyon lotor is a model species to test how habitat loss affects kin structure because, although relatively unaffected by habitat fragmentation (because of high dispersal ability), they are negatively affected by habitat loss (forest-related resources being important for female reproductive success). To elucidate the causes and consequences of kin structure in raccoons, we utilized genetic and demographic data collected from 998 individuals trapped from 27 spatially distinct habitat patches (local populations) situated in an agricultural ecosystem. Our results highlight an important, but often ignored fact: structural connectivity does not necessarily predict functional dispersal patterns in natural populations. Thus, in raccoons, local populations with low kin structure were associated with high landscape disturbance (i.e. high levels of habitat loss and low connectivity), and were characterized by demographic instability (i.e. high immigration, emigration and/or mortality). Alternatively, local populations exhibiting high kin structure were associated with low landscape disturbance and high demographic stability (i.e. high natality and philopatry). We propose that such increased philopatry in resource-rich patches could lead to a functional isolation (‘isolation by resource’) exacerbating the negative effects of landscape modification because of isolation by distance and/or resistance, especially in species with low dispersal capability. Our results also indicate that high levels of genetic diversity may be associated with low (rather than high) patch quality because populations in such patches could be composed predominantly of (unrelated) immigrants.