BACKGROUND/AIM: Propranolol is a beta-adrenergic receptor blocker which is used for the treatment of portal hypertension in patients with liver cirrhosis. The systemic exposure of propranolol may vary according to the extent of portal hypertension and liver function. The objective of this study was to propose a model for predicting the exposure of propranolol. METHODS: Thirty normal subjects, 18 patients with chronic active hepatitis (CAH), and 54 patients with cirrhosis were included in this study. Blood samples for pharmacokinetic analysis were taken up to 8 hours post-dose. Pharmacokinetic analysis was performed using a non-compartmental model. The extent of portal hypertension was evaluated by heart-to-liver radioactivity uptake ratio (H/L ratio) using 201TI per rectal scintigraphy. A multivariate regression analysis was performed to determine the best model for estimating the Cmax or AUC of propranolol. RESULTS: Twenty-four normal subjects, 18 CAH patients, and 36 cirrhosis patients completed the study. A multivariate stepwise linear regression analysis revealed that sex, weight, total bilirubin concentrations, platelet counts, and H/L ratio affected the exposure of propranolol: Cmax (ng/mL) = 50.976-18.743×sex[M:1;F:0]-0.408×weight+6.155×total bilirubin+35.328×H/L ratio (adjusted r2=0.440); AUClast (ng·h/mL) = 298.86–71.080×sex[M:1;F:0]–2.158×weight–0.312×platelet count+26.372×total bilirubin+176.745×H/L ratio (adjusted r2=0.500). CONCLUSION: A multivariate model based on laboratory tests, H/L ratio, body weight, and sex can predict the systemic exposure of propranolol and thereby inform the prescription of propranolol in patients with liver disease.