Nonalcoholic fatty liver disease (NAFLD) fibrosis score (NFS) has emerged as a useful predictor for a long‐term outcome in NAFLD patients. We evaluated the predictive performance of NFS for overall mortality in a Chinese population with NAFLD. All ultrasonography‐diagnosed NAFLD patients at our center between 1996 and 2011 were retrospectively recruited. The outcome was obtained by interview and the causes of death were confirmed by medical records. Area under receiver operating characteristic curve (AUC) was used to determine the predictive accuracy of NFS, BARD, FIB‐4 and aspartate aminotransferase/platelet ratio index (APRI) for mortality. Data from total of 180 eligible patients (median age 39 years; 96 males) were analyzed, with 12 deaths over a median follow‐up of 6.6 (range: 0.5‐14.8) ‐year. By Cox model analysis, NFS as a continuous variable other than 3 other scoring systems was identified as the only predictor for all‐cause mortality [hazard ratio 2.743, 95% confidential interval (CI) 1.670~4.504]. NFS yielded the highest AUC of 0.828 (95% CI 0.728~0.928, p<0.05), followed by FIB‐4, APRI and BARD with their AUCs of 0.806, 0.732 and 0.632 respectively (p<0.05, except for BARD). Collectively, NFS is a useful predictor of 6.6‐year all‐cause mortality for Chinese patients with NAFLD. This article is protected by copyright. All rights reserved.