Hepatitis C virus (HCV) infection or HCV-related liver diseases are now shown to cause more than 350,000 deaths every year. Adaptability of HCV genome to vary its composition and the existence of multiple strains makes it more difficult to combat the emergence of drug-resistant HCV infections. Among the HCV polyprotein which has both the structural and non-structural regions, the non-structural protein NS5B RNA-dependent RNA polymerase (RdRP) mainly mediates the catalytic role of RNA replication in conjunction with its viral protein machinery as well as host chaperone proteins. Lack of such RNA-dependent RNA polymerase enzyme in host had made it an attractive and hotly pursued target for drug discovery efforts. Recent drug discovery efforts targeting HCV RdRP have seen success with FDA approval for sofosbuvir as a direct-acting antiviral against HCV infection. However, variations in drug-binding sites induce drug resistance, and therefore targeting allosteric sites could delay the emergence of drug resistance. In this study, we focussed on allosteric thumb site II of the non-structural protein NS5B RNA-dependent RNA polymerase and developed a five-feature pharmacophore hypothesis/model which estimated the experimental activity with a strong correlation of 0.971 & 0.944 for training and test sets, respectively. Further, the Güner-Henry score of 0.6 suggests that the model was able to discern the active and inactive compounds and enrich the true positives during a database search. In this study, database search and molecular docking results supported by experimental HCV viral replication inhibition assays suggested ligands with best fitness to the pharmacophore model dock to the key residues involved in thumbs site II, which inhibited the HCV 1b viral replication in sub-micro-molar range. HCV nonstructural protein NS5B RNA-dependent RNA polymerase (RdRP) mediates the catalytic role of viral RNA replication. Lack of host RNA-dependent RNA polymerase enzyme had made it an attractive and hotly pursued target for drug discovery efforts. In this study, we developed a five-feature pharmacophore (3D QSAR) model for thumb site inhibitors of HCV RdRP, which estimated the experimental activity with a strong correlation of 0.971 & 0.944 for training and test sets, respectively. Our database search and molecular docking results suggested that the compounds 1 and 2 with best fitness to the pharmacophore model were predicted to interact with key residues involved in thumbs site II and could inhibit the HCV RdRP activity. Further, the compounds 1 and 2 potently inhibited HCV 1b viral replication in sub-micro-molar range.