Alzheimer's disease (AD) is a complex neurodegenerative disease whose pathological mechanisms involve dysregulation of oxidative stress and autophagy pathways. MAPK9, BAX and TFEB were used as key proteins. Wayne analysis was used to identify genes associated with autophagy and oxidative stress, and protein-protein interaction (PPI) networks were constructed to study the associations between key genes. The key genes were mined by machine learning algorithm and prognostic marker models were constructed. The immune characteristics of AD were investigated by gene collection enrichment analysis (GSEA) and immunoresponse pathway enrichment analysis, and the clinical application potential was evaluated by drug prediction and molecular docking analysis. Finally, Mendelian randomization (MR) analysis was used to verify the causal relationship between key genes and AD. The results showed that we successfully identified several genes associated with Alzheimer's disease, including MAPK9, BAX, and TFEB. GSEA analysis showed their active involvement in the immune response, indicating the importance of immune function in AD. Drug prediction models reveal potential therapeutic targets for these key genes.