Robust monitoring of wildlife populations to guide interventions is fundamental to conservation and wildlife management. Understanding how landscape characteristics are influencing recovering predator population dynamics is often vital to inform recovery strategies, management, and policy. The pine marten is recovering in the UK, however, spread and expansion has occurred at different rates across the country. We investigate how spatial variation in density of recovering pine marten populations is influenced by key habitat and human-related factors. Using genotyped non-invasively collected pine marten hair samples collected through standardised surveys covering four landscapes in Scotland over three seasons, we apply a multi-session, sex-structured, spatial capture-recapture model to estimate within- and between-landscapes spatial variation in density. We identify larger areas being used by individuals in recently recolonised areas compared to more established populations. We find significant differences in the spatial variation of density across landscapes ranging from 0.05 marten/km2 (CI: 0.03-0.07) in the recently colonised Borders region of southern Scotland to 0.21 (CI: 0.10-032) in the Trossachs region of central Scotland. There was substantial variation in marten density within landscapes. The amount of forest had a positive effect on population density, but not in the most recently recolonised population. Closer proximity to large gamebird shoots had a significant negative effect on marten density in one landscape suggesting that, despite being protected in law, martens suffer sufficiently elevated mortality close to some pheasant pens to locally depress their density. We did not find an effect of distance to nearest major road on spatial variation in population density in any landscape. While the mechanistic processes driving these patterns are not clear, our work highlights the importance of elucidating the processes underlying these effects of forest and land management.