The impact of the protein METTL3 on tumorigenesis is well-established in cancer research. It promotes cell growth, invasion, migration, and drug resistance. METTL3 is also involved in the modulation of hematopoietic stem cell differentiation. Inhibiting METTL3 presents a potential therapeutic strategy for myeloid leukemia. This study aimed to identify METTL3 inhibitors through a structure-based virtual screening approach, utilizing an in-house per-residue decomposition virtual screening protocol. We mapped the binding interaction profile of V22, a recognized METTL3 inhibitor, to construct a pharmacophore model for the systematic exploration of potential inhibitors within a chemical database. Four out of nine hit compounds retrieved from ZINC compounds database, showed promising results, and were further investigated. A comprehensive evaluation of the ADMET properties and physicochemical characteristics of these compounds revealed superior qualities compared to V22. Molecular dynamics (MD) trajectory analysis unveiled substantial structural conformational changes induced by these compounds within the METTL3 protein, offering potential insights into therapeutic inhibition. After mapping per-residue interaction footprints and examining toxicity profiles, we successfully identified the critical residues essential for activity and selectivity, informing our inhibitor design. Furthermore, the four compounds exhibited total binding energies of − 45.3 ± 3.3, − 40.1 ± 4.2, − 56.9 ± 3.3, and − 50.1 ± 4.1 kcal/mol for ZINC67367742, ZINC76585975, ZINC76603049, and ZINC76600653, respectively. The structural changes observed in proteins upon binding to specific compounds have important therapeutic implications. These alterations include increased stability, improved structural alignment, reduced flexibility, and greater compactness. These changes make these compounds promising candidates for developing more effective therapeutic inhibitors in the treatment of myeloid leukemia.