Alzheimer’s Disease (AD), the most common form of dementia, is driven by amyloid-beta plaques and tau tangles, yet few treatments address its underlying biochemistry. Apolipoprotein E4 (ApoE4), present in 13% of the population but over 50% of AD patients, is the strongest known risk factor. Despite differing in 1-2 residues in the N-terminal, ApoE isoforms diverge drastically: ApoE2 is protective, ApoE3 neutral, and ApoE4 pathogenic. This project builds on Tianhua Zhai’s identification of ApoE4 as a prime target, using drug discovery tools to design small molecules that modulate ApoE4 toward ApoE2 or ApoE3-like behavior. A computational pipeline was developed using diffusion modeling, a generative approach conditioned on protein pockets and scaffolds. Steps included docking known compounds on DiffDock-L to identify key pockets, generating over ten thousand molecules with Diffusion Structure-Based Drug Design, filtering candidates by Lipinski’s Rules, duplicates, size, and synthetic affinity through RDKit, and docking with CompassDock, DiffDock-L, and Boltz-2. Molecules were evaluated by minimized affinity, IC50 values, and druglikeness (QED). Ligand15_1 emerged as the strongest candidate, with significant affinity, low IC50, and a QED of 0.575. Molecule76_5 balanced strong affinity with low IC50 and acceptable druglikeness. Some ligands did not bind directly to residues R112 and R158 but still induced conformational changes that reduce harmful domain interactions. This pipeline can be applied to other ApoE4 domains, Alzheimer’s proteins such as BACE1, and cancer targets like ABCG2. While QED scores remain a challenge, wet-lab validation could confirm these results, paving the way for ApoE4 modulation in Alzheimer’s therapy.