α/β-Hydrolase domain-containing 6 (ABHD6) represents a potentially attractive therapeutic target for indirectly potentiating 2-arachidonoylglycerol signaling; however, the enzyme is currently largely uncharacterized. Here, we describe a five element, ligand-based pharmacophore model along with a refined homology model of ABHD6. Following a virtual screen of a modest database, both the pharmacophore and homology models were found to be highly predictive, preferentially identifying ABHD6 inhibitors over drug-like non-inhibitors. The models yield insight into the features required for optimal ligand binding to ABHD6 and the atomic structure of the binding site. In combination, the two models should be very helpful not only in high-throughput virtual screening, but also in lead optimization, and will facilitate the development of novel, selective ABHD6 inhibitors as potential drugs.