Diabetes is one of the top causes of death in the world with 425 million sufferers reported in 2017. About 90% of people with diabetes suffer from Type 2 Diabetes Mellitus (T2DM). Recent studies show that inhibiting the α-amylase enzyme can significantly decrease the postprandial blood glucose levels through blocking carbohydrate hydrolysis. Therefore, it can be a promising strategy for T2DM treatment. This research was aimed to find the new potential inhibitor for the a-amylase from lead-like compounds Molecular Operating Environment (MOE) database through fragmentbased drug design, combining with structure-based pharmacophore design method to obtain new drug candidate for T2DM. There were 653,214 lead-like compounds which were obtained from MOE database and screened based on the Astex Rules of Three along with toxicity filter to gain lead-like fragments. The filtered fragments were docked into the binding site of the a-amylase utilizing MOE 2014.09 software. Potential lead-like fragments were grown to generate 25,600 new ligands by utilizing DataWarrior v5.0.0 software, based on the Lipinski’s Rule of Five and toxicity filter. Molecular docking simulation and pharmacological tests were performed on the ligand libraries to acquire the best ligand, namely BGOJI, which were chosen according to the lowest ΔG binding score, RMSD value < 2, good molecular interaction, ADME/T-test result.