Observational studies have demonstrated a significant association between socio-economic status (SES) and Alzheimer's disease (AD). Nonetheless, the precise biological mechanisms underlying this association remain unclear. Therefore, we adopted a Mendelian Randomization (MR) approach to investigate the causal relationship between SES and genetic susceptibility to AD, as well as to explore the potential mediation effects of inflammation.Large-scale cohorts based on publicly available genome-wide association study (GWAS) datasets from European populations were employed for conducting the MR study. The primary criterion utilized was the inverse-variance weighting (IVW) model. Heterogeneity and horizontal pleiotropy were assessed. In addition, multivariate MR (MVMR) was utilized to correct the confounders. Moreover, a two-step MR approach was used to evaluate the potential mediating effects of factors on the causal effects between SES and AD.As indicated by the results of the IVW model, educational years (OR = 0.708, 95% CI 0.610-0.821, P < 0.001) and household income (OR = 0.746, 95% CI 0.566-0.982, P = 0.037) was associated with a decreased genetic susceptibility risk for AD. The univariable results showed that the causal effect of educational years on the lower risk of AD remained significant (OR = 0.643, 95% CI 0.467-0.886, P = 0.006). In addition, our findings indicated that C-reactive protein (CRP) played a role in the causal effect of educational years on AD. The proportions of mediation were - 50.08% (95% CI - 92.78; - 7.38%).These findings provided evidence supporting the causal effect of educational attainment lower AD risk, with inflammation playing a mediating role. These findings may inform prevention strategies and interventions directed toward AD. Future studies should explore other plausible biological mechanisms.© 2024. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany.