Simulation of virtually designed 20 compounds as COX-2 inhibitors using molecular modelling of protein–ligand interactions to predict drug structure–activity relationship was performed in this study. A synthetic route with a rational chemical approach to (E)-2-oxo-(thio)-4-substituted phenyl-6-styryl-1,2,3,4-tetrahydro-pyrimidine-5-caboxylic acid was designed and demonstrated. A comparative analysis of antimetabolite drug and corresponding metabolites (virtually designed compounds) provided a better understanding of rational drug design. COX-1(pdb entry: 1eqg) and COX-2(pdb entry: 6cox) enzymes docked with novel ligands were evaluated for binding energies. Lead optimization was performed by computational simulation: methoxy-substituted analogues displayed the highest negative ligand–protein-binding energies. These results prompted us to evaluate in vivo anti-inflammatory activity by carrageenan-induced paw oedema test in rats at a dose of 100 mg/kg. Ibuprofen was administered as standard drug. Lead compounds having significant activity were tested for in vitro cyclooxygenase isoenzyme inhibition assay and found to be more selective towards COX-2 as indicated by COX-2 selective index. The objective of our research is to accept the challenge of discovery of new drug. To ensure the desired target specificity and potency, bioavailability and lack of toxicity, our approach stems out lead generation from virtual screening to their synthesis and ends up with biological assays.