The factors influencing continuous renal replacement therapy (CRRT) duration for critically ill patients with acute kidney injury (AKI) are unclear. Therefore, we investigated the clinical factors that could influence the duration of CRRT for AKI survivors. In this retrospective observational study, the medical records of all hospital survivors who required CRRT for AKI in intensive care units were analyzed. The CRRT duration (median, 6 days) was categorized as short-duration CRRT (≤ 6 days, n = 65) and long-duration CRRT (> 6 days, n = 59), according to the median CRRT duration. A urine output of less than 0.5 mL/kg/h (adjusted odds ratio [OR], 3.4; P = 0.010), mechanical ventilation use (adjusted OR, 7.9; P = 0.001), and extracorporeal membrane oxygenation (ECMO) use (adjusted OR, 6.5; P = 0.010) were independent predictors of long-duration CRRT, whereas serum creatinine and neutrophil gelatinase-associated lipocalin were not significant predictors. A clinical model demonstrated a good discriminatory ability to predict long-duration CRRT (area under the curve, 0.84; 95% confidence interval, 0.76-0.90). The urine output immediately before CRRT initiation and factors associated with disease severity significantly affected the duration of CRRT. Simultaneously considering the urine output, mechanical ventilation use, and ECMO use predicted CRRT duration in AKI survivors.