Green infrastructure (GI) is a strategically planned network of natural and semi-natural areas designed and managed to sustain biodiversity and deliver ecosystem services. To support GI planning for boreal regions, we mapped structurally mature forest habitat for the Siberian jay (Perisoreus infaustus), a taiga species often portrayed as an old-growth umbrella species associated with late successional spruce-dominated conifer forest. Using forest data from Sweden, we developed a new habitat suitability model based on airborne laser-scanning (LiDAR) canopy-structure metrics to test whether this commonly assumed association holds across gradients of forest management and proximity to human activity. We compared LiDAR-based predictions with those from a pre-existing Heureka stand-age model, validated both models using occurrence data, and used the resulting suitability maps to quantify habitat patch size and evaluate the influence of human presence. The LiDAR model estimated 44% of the suitable habitat area predicted by the Heureka model and identified the largest high-suitable blocks mainly in the southern part of the study area, dominated by older multilayer pine stands. In both models, predicted suitable habitat increased with occurrence class, and mean annual presence was lower near settlements, indicating a negative effect of human settlements associated with corvid nest predators. Suitable habitat was highly fragmented. Most patches were < 1 ha, while patches > 1000 ha were rare. Forest management should prioritize retention of older multilayer forest patches, and connectivity of remaining habitat, especially small patches that due to edge effect are more vulnerable to external disturbances.