Postoperative cognitive dysfunction (POCD) is a frequent complication of cardiac surgery. The present study aimed to quantify changes in Mini‑Mental State Examination (MMSE) scores following on‑pump cardiac surgery and explore potential study‑level moderators associated with variability in cognitive outcomes. A systematic review and random‑effects meta‑analysis was performed following PRISMA guidelines. MEDLINE, Embase, Cumulative Index to Nursing & Allied Health Literature, Cochrane, PsycINFO, Web of Science, and Scopus were searched up to May 2025; eight prospective single‑center cohorts (867 patients) met predefined criteria, providing 14 independent pre/postoperative comparisons of cognitive function. The mean MMSE score was extracted at baseline and final reported follow‑up (7 days to 6 months). Hedges’ g was pooled using a random‑effects model; heterogeneity was summarized with τ² and I², and a 95% prediction interval (PI) was calculated. Overall, cardiac surgery was associated with a moderate decline in MMSE scores (g=‑0.60, 95% confidence interval, ‑0.85 to ‑0.35; P<0.001). Heterogeneity was high (τ²=0.178; I²=88.8%), yet leave‑one‑out and Hartung‑Knapp adjustments produced similar estimates. The 95% PI (‑1.56 to 0.36) spans both negative and slightly positive values, indicating that while most cohorts show cognitive decline, the inclusion of values above zero suggests that small improvements remain possible in some populations. Subgroup analysis showed a large decline among patients meeting POCD criteria (g=‑0.893) compared with a small, non‑significant change in non‑POCD patients (g=‑0.274; Q_between=9.08, P=0.003). Meta‑regression identified cardiopulmonary‑bypass time as the only statistically significant moderator (β=‑0.014 MMSE SD units/min; P=0.023), explaining ~50% of between‑study variance. Funnel‑plot tests suggested possible small‑study effects, but trim‑and‑fill imputed no missing studies and Rosenthal’s fail‑safe N was 431, supporting the robustness of the pooled estimate. These findings suggest that on‑pump cardiac surgery is frequently associated with a moderate decline in global cognitive performance as measured by the MMSE, particularly among patients who develop clinical POCD. Awareness of this association may support perioperative cognitive monitoring and efforts to minimize modifiable intraoperative factors such as prolonged cardiopulmonary bypass duration.