PURPOSE: For patients with advanced or unresectable hepatocellular carcinoma (HCC), safe and effective therapies are urgently needed to improve their long-term prognosis. Although the guidelines recommend first-line treatments such as sorafenib, lenvatinib, and atezolizumab in combination with bevacizumab (T+A) and second-line treatments such as regorafenib, the efficacy comparison between drugs is lacking, that is, a treatment is not recommended as the optimal or alternative choice for a specific patient population. Therefore, we will conduct a high-quality network meta-analysis based on Phase III randomized controlled trials (RCTs) to systematically evaluate and compare overall survival (OS), progression-free survival (PFS), objective response rate (ORR), and serious adverse events (SAE) of different treatment protocols in the context of first-line and second-line therapies, which are critical for clinical decision making and prognostic improvement in advanced HCC patients. METHODS: The studies of interest were Phase III RCTs evaluating the efficacy or safety of first- or second-line therapies in patients with unresectable or advanced HCC. Literature published in English from the four databases of PubMed, Embase, Cochrane Library, and Web of Science was comprehensively searched from the inception to May 23, 2022. Outcomes of interest included OS, PFS, ORR, and SAE. A league table was developed to show the results of the comparison between different treatments. A histogram of cumulative probability was drawn to discuss the ranking probability of treatments based on different outcomes. The effectiveness and safety of various treatments were comprehensively considered and the two-dimensional diagram was plotted to guide clinical practice. The Gemtc package in R Studio was used for network meta-analysis in a Bayesian framework. RESULTS: The results showed that HAIC-FO was superior to T+A regimen, regardless of OS, PFS or ORR. TACE combined with lenvatinib performed better than T+A in PFS, and ORR. In addition to the T+A regimen, Sintilimab combined with IBI305 and camrelizumab combined with apatinib were also associated with longer OS, PFS, and ORR, and their SAE incidence was not higher than that of T+A, especially for camrelizumab combined with apatinib, its safety was better than that of T+A regimen. There were no new treatments or combinations that were more effective than regorafenib. It was important to note that for PFS, the efficacy of apatinib and cabozantinib was not statistically different from that of regorafenib, so these two treatments could be used as alternative treatment options in cases where regorafenib was not tolerated or treatment failed. CONCLUSIONS: We conducted a network meta-analysis to evaluate the efficacy and safety of multiple treatment modalities by integrating the results of direct and indirect comparisons. This study included high-quality multicenter Phase III RCTs, collated and summarized all treatments involved in advanced or unresectable HCC in first-line and second-line settings, and compared with T+A and regorafenib, respectively, and ranked based on efficacy and safety to support clinical decision making.