Infection by the mosquito-borne dengue virus causes dengue fever and the sometimes fatal dengue hemorrhagic fever. The increasing number of dengue infections per year suggests that the virus is becoming more virulent and its transmission is expanding. Nevertheless, no effective treatment for dengue infection currently exists. In a search for antiviral agents effective against dengue virus, we investigated the potential of targeting a structural protein site rather than an enzymatic one. Using this approach, we now report the discovery of a small molecule ligand that inhibits viral growth. Our results also provide the first evidence that the binding site, a pocket located at the hinge between domains 1 and 2 of the envelope protein (E protein) on the virus surface, is a valid target for antiviral therapy. Ligand candidates were identified from libraries of approximately 142,000 compounds using a computational high-throughput screening protocol targeting this pocket of the E protein. Cell-based assays were conducted on 23 top-ranked compounds. Among four with good antiviral activity profiles, the compound P02 was found to inhibit viral reproduction at micromolar concentrations. Using saturation transfer difference NMR spectroscopy, we also show that the compound binds virus and competes for binding E protein with the known ligand N-octyl-beta-D-glucoside. Together, the results are consistent with an inhibition mechanism against maturation or host-cell entry mediated by ligand binding to the E-protein pocket. P02 is a promising lead compound for future development of an effective treatment against dengue virus and related flaviviruses.