The β-amyloid precursor protein cleavage enzyme (BACE) has been conceived to be an attractive therapeutic target to control Alzheimer's disease (AD). Validated ligand-based pharmacophore mapping was combined with 3D QSAR modeling approaches that include CoMFA, CoMSIA and HQSAR techniques to identify structural and physico-chemical requirements for a potential BACE inhibitor using a database containing 980 structurally diverse compounds, assembled from different reports. A structure-based docking technique was also used to validate the features obtained from the ligand-based models, which were further used to screen the database of compounds designed by a de novo approach. Contour maps of 3D QSAR models, CoMFA (R² = 0.880, se = 0.402, Q² = 0.596, R_pred² = 0.713,) and CoMSIA (R² = 0.903, se = 0.362, Q² = 0.578, R_pred² = 0.715), and a pharmacophore space model (R² = 0.833, rmsd = 1.578, Q² = 0.845, R_pred² = 0. 764) depict that the models are robust and provide an explanation of the important features such as steric, electrostatic, hydrophobic, positive ionization, hydrogen bond acceptor and donor properties, which play important roles for interaction with the receptor site cavity. The HQSAR study (R² = 0.823, se = 0.488, Q² = 0.823, R_pred² = 0.768) and de novo design, which generate new fragments, illustrated the important molecular fingerprints for inhibition. The docking study elucidated the important interactions between the amino acid residues (Gly11, Thr72, Asp228, Gly230, Thr231, Arg235) at the catalytic site of the receptor and the ligand, indicating the structural requirements of the inhibitors. The de novo designed molecules were further screened for ADMET properties, and ligand–receptor interactions of the top hits were analysed by molecular docking to explore pharmacophore features of BACE inhibitors.