Oral squamous cell carcinoma (OSCC) remains an important cause of cancer morbidity and mortality, and new multi-target approaches are required to increase therapeutic options. Drug repurposing offers a promising strategy to identify new anticancer applications for clinically approved drugs with established safety profiles. An integrative in silico workflow was designed to identify overlapping targets between selected antifungal drugs and OSCC-associated genes based on network pharmacology and protein-protein interaction analysis. Network pharmacology identified 144 common targets for antifungals and OSCC, and EGFR and MMP9 were selected as druggable hubs based on their contributions to cancer progression. Molecular Docking revealed promising binding, with fluconazole (−8.5 kcal/mol) and itraconazole (−8.2 kcal/mol) showing better binding than the reference inhibitors tanomastat (−8.3) and erlotinib (−6.9) against MMP9 and EGFR, respectively. However, 200 ns molecular dynamics simulations and MM-PBSA calculations showed a different reality. Erlotinib and tanomastat showed the dynamic signature of specific binding, including persistent hydrogen bonds, restricted PCA clusters, and deep free energy minima. In contrast, fluconazole-EGFR showed thermodynamically unfavorable binding (+20.249 kJ/mol), fluconazole-MMP9 showed transient contacts mediated through zinc, and itraconazole-EGFR did not form persistent hydrogen bonds. Notably, the highest RMSD (0.31 ± 0.11 nm), protein destabilisation and unreliable binding energy (±101.005 kJ/mol) were shown by itraconazole-MMP9. The results imply that fluconazole and itraconazole are not specific and direct inhibitors of EGFR or MMP9, and their docking scores are artifacts from zinc chelation and hydrophobic overestimation. This study provides a cautionary computational framework emphasizing that rigorous MD and MM-PBSA validation is indispensable for reliable drug repurposing.