Baobab, Adansonia digitata L. is a widespread multipurpose savannah tree in Africa. Although previous studies highlighted the potential impacts of climate on the species, we still lack sufficient understanding of how regional variation in ecological tolerance might influence its future resilience across the continent. This study used ensemble species distribution models (SDMs) to forecast suitable habitats for A. digitata under two climate scenarios (SSP 245 and SSP 585) for 2050 and 2070, across four climate zones: humid, subhumid, semi-arid, and arid. We built both whole-species and climate-zone-specific models to assess the role of niche differentiation on projected habitat suitability. Isothermality (36.75 %) was the most influential variable in the whole-species model, while soil properties (26.41 %) constrained habitat suitability. In zonespecific models, key drivers changed with the zone: precipitation of the driest month and soil properties dominated in the subhumid zone; annual precipitation and soil properties in the semiarid; maximum temperature and annual precipitation in the humid; and mean diurnal temperature range in the arid. Niche overlap between zones was low (Schoener's D = 0.06 - 0.13), suggesting a broad climate tolerance. Whole-species models predicted slight habitat loss (0.5 - 1.12 %) by 2070, especially under SSP 585, whereas zone-specific models showed increases (1.48 - 3.13 %), irrespective of the scenarios and horizons, particularly in the humid zone. These findings underscore the importance of accounting for regional variation in SDMs for widespread species. Baobab's broad climatic tolerance may support resilience to climate change, though reciprocal transplant experiments are needed to disentangle the role of local adaptation from plasticity.