The desert night lizard (Xantusia vigilis) is a habitat specialist abundantly spread across arid regions of the North American southwest, often reliant on fallen Joshua tree branches (Yucca brevifolia) for shelter. Future climate change impacts on Y. brevifolia are therefore of particular concern for the lizard’s ecological conservation. Here, we explored the impacts of climate change on X. vigilis’ geographical distribution using a Species Distribution Model (SDM) of its climatic niche, additionally aiming to relate these responses to those for Y. brevifolia. We trained the SDM with a set of ten uncorrelated WorldClim Bioclimatic variables (1970-2000 averages) and presences of each species (>1000 unique locations). A random forest classifier performed best from a set of over ten candidates (including Maxent), emerging as the best predictive model of the current geographic distribution (e.g., OOB misclassification error ~ 2-3%). We then projected the SDM to future climate conditions, simulated with eight climate models from CMIP6 over four Shared Socioeconomic Pathways, for the years 2040-2100. Under these scenarios the range of X. vigilis was predicted to decline to between 11% to 55% of its current range, assuming little or no dispersal. In addition, a single climate model, CanESM5, consistently predicted the most dire scenario of future habitat suitability. We also measured overlap with projections of Y. brevifolia’s distribution. Our results highlight the importance of including symbiotic and other ecologically important species into models of climate change effects on geographic distributions, with conservation risks possibly heightened for localities which face extreme climate events, such as wildfires.