Background Assessing hepatic functional reserve before hepatectomy is beneficial to reduce the incidence of posthepatectomy liver failure (PHLF). This study aimed to compare the ability of the Child‐Pugh score, model for end‐stage liver disease (MELD) score, and retention test at 15 minutes (indocyanine green ICG‐R15) to assess hepatic functional reserve. Methods A total of 185 patients with hepatocellular carcinoma (HCC) undergoing hepatectomy were enrolled in this study. The ability of Child‐Pugh score, MELD score, and ICG‐R15 predicting severe PHLF were compared. Results A total of 23 patients (12.4%) developed severe PHLF. Multivariate analyses identified that platelet count, ICG‐R15, clinically significant portal hypertension, and major resection were independent factors for predicting severe PHLF. The area under the receiver operating characteristic curve of ICG‐R15 for predicting severe PHLF was higher than that of both Child‐Pugh score and MELD score. With an optimal cutoff value of 7.1%, the sensitivity and specificity of ICG‐R15 for predicting severe PHLF were 52.2% and 89.5%, respectively. Both the incidence of severe PHLF and mortality in patients with ICG‐R15 >7.1% were significantly higher than the figures for patients with ICG‐R15 ≤7.1%. Conclusion ICG‐R15 is more accurate than the Child‐Pugh score and MELD score in predicting hepatic functional reserve before hepatectomy.