Background & aims: Sepsis-associated acute kidney injury (SA-AKI) is associated with high mortality rates and lacks effective interventions. Omega-3 polyunsaturated fatty acids, abundant in fish oil, possess anti-inflammatory and immune-regulatory properties; however, their impact on the prognosis of SA-AKI remains unclear. This study aims to explore the association between fish oil supplementation and clinical outcomes in critically ill patients with SA-AKI. Method: A retrospective cohort study was conducted using the MIMIC-IV database. A total of 14,234 patients with SA-AKI were included, with 142 patients in the fish oil supplementation group and 142 in the control group, matched by propensity scores. Cox regression, logistic regression, and linear regression models were employed to evaluate the impact of fish oil supplementation on survival, renal function recovery, and length of hospital stay. Sensitivity analyses were performed to assess the robustness of the results. Results: Survival analysis suggested that the fish oil supplementation group had a higher 21-day survival rate compared to the control group. Cox regression indicated that fish oil supplementation was associated with lower 21-day mortality (HR = 0.74, 95 % CI: 0.59-0.99, P = 0.032). Logistic regression also showed that it was associated with improved renal function recovery (OR = 1.76, 95 % CI: 1.03-3.01, P = 0.040). Subgroup analysis suggested that the effect of fish oil supplementation may vary based on the presence of congestive heart failure and CRRT/RRT treatment status. No significant difference was observed in ICU length of stay between the two groups. Conclusion: Fish oil supplementation in critically ill patients with SA-AKI was associated with reduced 21-day mortality and improved renal function recovery. These findings support the need for prospective randomized controlled trials to confirm the therapeutic potential of omega-3 PUFA in sepsis-related organ dysfunction. (c) 2026 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.