We delineated the existing empirical ranges of western and eastern Joshua trees (Yucca brevifolia and Y. jaegeriana, respectively) with high fidelity across their ranges in Arizona, California, Nevada, and Utah, USA. Most species distribution models (SDMs) rely on sparse species occurrence datasets and random pseudoabsences. In contrast, the tall stature and distinctive branching arms of Joshua trees enabled us to definitively identify this species in publicly available satellite imagery, allowing us to use intensive visual grid searches to map empirical presences and absences at a 0.25 km2 resolution across most of the species’ ranges. We used the resulting presence/absence data to train species distribution models (SDMs) for each Joshua tree species, as well as a rangewide model comprising the distribution data from both species. Species distribution models link species' presence / absence data with environmental characteristics including topography, climate, and soils, revealing the contemporary environmental gradients (during the past 40 years) with greatest influence on the current habitat of adult trees. Each SDM in this data release consists of a GeoTiff raster file giving the probability of occurrence for one or both species of Joshua tree, with values ranging from 0 (very low probability of species occurrence) to 1 (very high probability of species occurrence). SDM raster layers are named according to the species they represent.