Background <p>Multiple staging systems for hepatocellular carcinoma (HCC) have been proposed, where Child–Pugh class (CP) is utilized to assess liver function. However, several inherent limitations occur in CP. We investigated whether replacement of CP by model for end-stage liver disease grade (MELD) or albumin–bilirubin grade (ALBI) in currently used HCC staging systems could achieve better prediction performance.</p> Methods <p>568 patients first diagnosed with HCC were retrospectively analyzed. We compared each original and modified systems by calculating their Harrell’s concordance index (C-index), Wald χ², and Corrected Akaike information criterion (AICc) as well as plotting decision curves and calibration curves by R version 4.3.1.</p> Results <p>The study identified severity of liver dysfunction, malignancy of tumor, and health status of patients as crucial factors of prognosis in HCC. In the entire cohort, replacement of CP by ALBI in staging systems resulted in comparable or even improved prediction performance for HCC prognosis, with higher C-index, higher Wald χ², and lower AICc, while incorporation of MELD in staging systems failed to do so. Similar findings were observed in the subgroups when patients were stratified according to different etiologies (hepatitis B virus infection or cirrhosis) and diverse therapy strategies (curative or non-curative treatments). Notably, ALBI-based Hong Kong Liver Cancer staging system was the optimal prognostic model with superior outcome prediction in different cohorts (the entire cohort: C-index = 0.776; Wald χ² = 241.8; AICc = 2469.079).</p> Conclusion <p>Our study confirms comparable or, in some cases, superior prognostic performance of the ALBI grade to the CP class across specific HCC staging systems. ALBI may serve as a complementary or alternative measure that may enhance prognostic accuracy, conducive to therapeutic decisions of oncologists and to the effective management of HCC patients.</p>