Species distribution models have become increasingly important tools for species conservation. This modeling approach can help guide conservation practitioners and inform decisions. Distribution models are particularly relevant for rare species, whose habitat associations are often not well understood. Using species occurrence data, and a variety of predictor variables, we developed three individual distribution models and a weighted ensemble model for the Puerto Rican harlequin butterfly (Atlantea tulita). The ensemble model had the greatest accuracy (AUC = 0.92). Further, the ensemble model indicated 7.1% of the main island of Puerto Rico encompassed suitable habitat for the harlequin butterfly. However, only 0.5% was classified as including the greatest suitability. Using an ensemble modeling approach to delineate areas of the island with suitable environmental conditions may improve habitat conservation efforts for the species.