To investigate the association between long-term changes of serum total bile acid and hepatocellular carcinoma in chronic hepatitis B patients, we did a retrospective cohort study of 2262 chronic hepatitis B patients with regular antiviral treatment using data from the Hepatitis Biobank at Southwest Hospital Program from 2004 to 2014. Patients in the study were classified into 3 groups according to persistence of elevated serum total bile acid during follow-up: none-low, medium, and high persistence of elevated serum total bile acid. The association between persistence of elevated serum total bile acid and hepatocellular carcinoma was estimated using Cox proportional hazard models and Kaplan-Meier analysis including information about patients' demographic and clinical characteristics. There were 62 hepatocellular carcinoma cases during a total follow-up of 14756.5 person-years in the retrospective study. Compared to patients with none-low persistence of elevated total bile acid, the multivariate adjusted hazard ratios (95% confidence interval) were 2.37 (1.16-4.84), and 2.57 (1.28-5.16) for patients with medium, and high persistence of elevated total bile acid. Our findings identified persistence of elevated serum total bile acid as an independent risk factor of hepatocellular carcinoma in chronic hepatitis B patients.