The cannabinoid CB1 and CB2 receptors belong to the super family of G protein-coupled receptors. Antagonists for each receptor sub-type, as well as four structural classes of agonists that bind to both receptors, have been identified. In the absence of an experimentally determined structure for each of these two receptors, computational molecular modeling approaches have been employed to begin to probe structure-function relationships. Molecular modeling studies have been approached from two perspectives: calculations involving only ligands (Ligand-Ligand Approach) or calculations of the interaction of a ligand with its receptor macromolecule (Ligand-Receptor Approach) [49]. The Ligand-Ligand Approach does not directly consider the structure of the ligand binding site, but attempts to infer information about this site from a correlation between experimentally determined biological activities and the structural and electronic features of a series of small molecules. Ligand-Ligand Approaches result in development of pharmacophore models. Although closer to the event of interest, the study of the binding of a ligand to its receptor is less common because it requires a working knowledge of the receptor structure [54]. Mutation/ chimera studies, as well as structure activity relationships can be used to test models developed in a Ligand-Receptor Approach. This review considers both Ligand-Ligand and Ligand-Receptor computational studies of the CB1 and CB2 receptors. Challenges for further modeling studies are also identified.