Sorafenib may improve progression-free survival (PFS) and overall survival (OS) of advanced hepatocellular carcinoma (HCC). However, the survival benefit is short lived and survivals after progressive disease (PD) have not been well characterized. This study aimed to evaluate the survival predictors of OS and postprogression survival (PPS) in advanced HCC patients receiving sorafenib treatment. Consecutive 149 HCC patients receiving sorafenib under National Health Insurance were retrospectively enrolled. All patients fulfilled the reimbursement criteria: Barcelona Clinic Liver Cancer stage C HCC with macroscopic vascular invasion or extrahepatic metastasis (Mets), and Child-Pugh class A. Radiologic assessment was performed at a 2-month interval using modified Response Evaluation Criteria in Solid Tumors. Patients who maintained Eastern Cooperative Oncology Group ≤2 and Child-Pugh class A at PD were assumed to be candidates for second-line treatment. During the median follow-up period of 7.5 months (range, 1.1-18.5), PD developed in 120 (80.5%) patients and 96 (64.4%) deaths occurred. The median PFS, OS, and PPS were 2.5, 8.0, and 4.6 months, respectively. In general, patients with Mets only had better OS and PPS than those with portal vein invasion. Independent predictors of OS include baseline performance status (hazard ratio [HR] = 1.956), tumor size (HR = 1.597), alpha-fetoprotein (HR = 1.869), discontinuation of sorafenib due to liver function deterioration (LD) (HR = 6.142), or concurrent PD and LD (HR = 2.661) and PD within 4 months (HR = 5.164). Independent predictors of PPS include deteriorated performance status (HR = 7.680), deteriorated liver functions (HR = 5.603), bilirubin (HR = 2.114), early PD (HR = 6.109), and new extrahepatic lesion (HR = 1.804). In 46 candidates for second-line trials, development of new extrahepatic lesion independently predicts poorer PPS (HR = 3.669). In conclusion performance status, liver functions, early disease progression, and progression pattern are important determinants of survival after sorafenib failure. These factors should be considered in clinical practice and second-line trial designs for patients with sorafenib failure.