OBJECTIVE The hepatic venous pressure gradient (HVPG) has been employed as the gold standard for indicating the portal venous pressure gradient (PPG) in the diagnosis of portal hypertension (PHT). However, little has been reported on whether the HVPG can accurately estimate the PPG in patients with hepatic vein collateral shunts. We aimed to explore the correlation between the HVPG and the PPG in hepatitis B cirrhosis patients with different hepatic vein anatomies. METHODS A total of 461 hepatitis B cirrhosis patients with portal hypertension (PHT) who were treated with a transjugular intrahepatic portosystemic shunt (TIPS) between January 2016 and June 2020 were included. All patients underwent various venous pressure measurements and balloon-occluded compression hepatic venography during the TIPS operation. Agreements were evaluated by Pearson's correlation and the Bland-Altman method. Disagreements were assessed by paired t tests. RESULTS The correlation coefficient (r) values (P < 0.001) between the HVPG and the PPG of the early (151 patients, 32.8 %), middle (73 patients, 15.8 %), late (46 patients, 10.0 %), portal vein (151 patients, 32.8 %), and no lateral branch development groups (40 patients, 8.7 %) were 0.373, 0.487, 0.569, 0.690, and 0.575, respectively; the determination coefficient (R2) values were 0.139, 0.238, 0.323, 0.475, and 0.330, respectively. According to the Bland-Altman method, agreement was the greatest in the portal vein development group, with the 95 % limits of agreement (95 % LoA, mean differences ± 1.96 SD) being the smallest. The differences were statistically significant (P < 0.05). CONCLUSIONS The correlation between the HVPG and the PPG is the worst in early lateral branch development, followed by middle development, and the influence of lateral branches becomes significantly reduced in late development. Hepatic venous collateral formation is a vital factor for underestimation of the HVPG, which is the most accurate predictor of PPG in patients with portal vein development. Patients with no collateral channel development in the hepatic vein have a higher HVPG than PPG, which is an important reason for overestimation of the HVPG.