For an adaptation policy to be effective, it is essential to divide species populations into sub-populations, especially for species with diverse environmental conditions. In this study, occurrence data for Adansonia digitata L., a multi-purpose tree species, were collected from the Global Biodiversity Information Fund portal (1513), fieldwork (421), and other studies (421), and other studies (195) across West Africa. Principal component analysis was used to divide the sample into four subpopulations based on 19 bioclimatic variables, altitude, and soil data. Using a stack species distribution modeling (SSDM) approach, nine modeling algorithms were created for each subpopulation and for the overall sample. The results showed that the distribution of the subpopulations depended on the latitude and that the niches were differentiated and did not overlap. Environmental variables contributed differently across subpopulations. Overall, it was found that 47.63% of West Africa was not suitable for the cultivation and conservation of the species, while 15.38% was very suitable. The most favorable countries are Burkina Faso, Benin, Benin, Senegal and Togo.