The dopamine 3 (D3) subtype receptor has been implicated in several neurological conditions, and potent and selective D3 ligands may have therapeutic potential for the treatment of drug addiction, Parkinson's disease, and schizophrenia. In this paper, we report computational homology modeling of the D3 receptor based upon the high-resolution X-ray structure of rhodopsin, extensive structural refinement in the presence of explicit lipid bilayer and water environment, and validation of the refined D3 structural models using experimental data. We further describe the development, validation, and application of a hybrid computational screening approach for the discovery of several classes of novel and potent D3 ligands. This computational approach employs stepwise pharmacophore and structure-based searching of a large three-dimensional chemical database for the identification of potential D3 ligands. The obtained hits are then subjected to structural novelty screening, and the most promising compounds are tested in a D3 binding assay. Using this approach we identified four compounds with K(i) values better than 100 nM and eight compounds with K(i) values better than 1 microM out of 20 compounds selected for testing in the D3 receptor binding assay. Our results suggest that the D3 structural models obtained from this study may be useful for the discovery and design of novel and potent D3 ligands. Furthermore, the employed hybrid approach may be more effective for lead discovery from a large chemical database than either pharmacophore-based or structure-based database screening alone.