Background: The life expectancy of patients with advanced hepatocellular carcinoma (HCC) enrolled in clinical trials should be at least 3 month to test drug response. However, the prognosis of these patients was poor with median overall survival (OS) shorter than 3 months. This study was aimed to evaluate different prognosis score systems for HCC in predicting prognosis and 3-month OS of patients with advanced HCC. Methods: From November 2008 to April 2010, a total of 208 patients with advanced HCC who were not amendable to locoregional therapy were included in this study. Data were collected to classify patients according to Japan Integrated Staging scoring system, TNM stage by liver Cancer Study Group of Japan criteria, National Comprehensive Cancer Network Stage system, Cancer of the liver Italian Program scoring system for hepatocelluar carcinoma (CLIP), Advanced Liver Cancer Prognostic System (ALCPS), model of end-stage liver disease, French scoring system, Chinese University Prognostic Index staging system for HCC (CUPI), OKUDA score system, Tokyo score system, Barcelona Clinic Liver Cancer staging for hepatocelluar carcinoma at diagnosis. OS and 3-month OS were the end points used in the analysis. Survival analysis, Cox regression, relative operating characteristic (ROC), and logistic regression were utilized to access the prognostic value of each score system. Results: ROC analysis confirmed that ALCPS was the best prognostic system for 3-month OS. The cut-off point for ALCPS was 14 with sensitivity of 81.82% and specificity of 79.01%, which was reinforced by the survival variation on the two subgroups divided by ROC analysis generated cutoff point. CLIP and CUPI presented no statistical difference in prognostic discrimination with ALCPS, however, with relatively lower accuracy. Conclusions: ALCPS was the best score system in prediction of OS and 3-month OS among the 12 systems analyzed and a proper inclusion criteria for clinical trials for Chinese advanced HCC patients.