This dataset includes raster species distribution models of five pine species in the Madrean sky islands of México and the United States. Models were trained using a large bi-national field dataset (n = 1,416) of pine species observations. Species distribution models were created for the following pine species: P. arizonica (syn. P. ponderosa var. arizonica); P. discolor (syn. P. cembroides var. bicolor); P. chihuahuana (syn. P. leiophylla var. chihuahuana). We generated 2 distribution models for each species, one model is based on topographic and bioclimatic variables (1km resolution), and the second based on topographic and Landsat remote sensing variables (30m resolution). Naming convention follows a 6 letter species code based on the latin name (i.e. Pinus arizonica = pinari) followed by f the model type (topo = topographic, clim = bioclimatic, and ndvi30 = Landsat 30m NDVI variables). A topographic and bioclimatic species distribution model for Pinus arizonica is therefore "pinari.topo-clim.tiff". Species distribution rasters reflect probability of species occurrence (0-low, 1-high).