Purpose <p>Using Mendelian Randomization (MR) analysis, this study aimed to assess any causal effect of educational attainment on the risk of delirium and to determine whether smoking mediates this association.</p> Methods <p>We obtained genome-wide association study (GWAS) summary-level statistics for educational attainment (<i>n</i> = 765,283), age of smoking initiation (<i>n</i> = 341,427), cigarettes per day (<i>n</i> = 337,334) and delirium (4,381 cases, 469,981 controls) from GWAS repositories. The inverse variance weighted approach served as the main analytical strategy for causal estimation. Multiple sensitivity tests were used to assess the robustness of MR analyses. We evaluated the causal effect of educational attainment on delirium, the effect of educational attainment on smoking-related traits and the effect of smoking-related traits on delirium. We then performed mediation analysis to evaluate the mediating effect of smoking traits on the association between educational attainment and delirium. In addition, we performed linkage disequilibrium score regression (LDSC) to evaluate genetic correlations between traits.</p> Results <p>Higher educational attainment was significantly associated with a lower delirium risk (OR = 0.767, 95% CI: 0.637–0.922, <i>P</i> = 0.005). Age of smoking initiation was positively associated with educational attainment (<i>β</i> = 0.289, <i>P</i> = 8.89 × 10^-133) and inversely associated with delirium risk (OR = 0.553, <i>P</i> = 0.014), whereas cigarettes per day was inversely associated with educational attainment (<i>β</i> = –0.315, <i>P</i> = 1.70 × 10^-32) and positively associated with delirium risk (OR = 1.238, <i>P</i> = 0.022). Sensitivity analyses indicated that the MR results were not affected by heterogeneity or horizontal pleiotropy. Mediation analysis indicated that the age of smoking initiation and cigarettes per day mediated 64.4% (<i>P</i> = 0.014) and 25.3% (<i>P</i> = 0.024) of the total effect, respectively. LDSC analysis revealed a significant negative genetic correlation between educational attainment and delirium, and between age of smoking initiation and delirium. Educational attainment also showed strong correlations with smoking traits, positively with age of initiation and negatively with cigarettes per day.</p> Conclusion <p>The findings of this MR study support the notion that higher educational attainment may reduce the risk of delirium by promoting a later age of smoking initiation and decreasing the number of cigarettes smoked per day.</p>