It has been nearly 50 years since Gaddum and Picarelli characterized serotonin M-receptors using an isolated guinea pig ileum preparation; M-receptors, known today as 5-HT3 receptors, are just one of about fifteen different types of serotonin receptors. The discovery of this new class of serotonin receptors evoked a necessity for new ligands to study their pharmacological activity. Today, several 5-HT3 ligands are employed therapeutically in the treatment of emesis, with further potential for the treatment of migraine and numerous central nervous disorders. Using radioligand binding techniques several chemical classes of compounds have been identified and investigated. These studies led to formulation of structure-affinity relationships (SAFIR) and the first two-dimensional pharmacophore model for 5-HT3 receptor ligand binding. Since then, several new SAFIR studies have been conducted and new, more advanced three-dimensional pharmacophore models have been proposed using molecular modeling techniques. In diverse functional assays, the majority of reported 5-HT3 derivatives that meet the established pharmacophore model requirements showed antagonist activity. Structure-activity relationships (SAR) for various classes of 5-HT3 receptor antagonists have been developed. The introduction of m-chlorophenylbiguanide (m-CPBG), a reasonably selective 5-HT3 agonist, called for redefinition of existing pharmacophores. In general, 5-HT3 agonists failed to adhere to 5-HT3 pharmacophore models. This triggered new studies leading to formulation of SAFIR and SAR for agents with 5-HT3 receptor agonist activity. Progress in the development of 5-HT3 agonists and proposed agonist pharmacophore models are reviewed.