Objectives. Recent epidemiological studies suggest that metformin treatment may reduce the risks of cancer and overall cancer mortality among patients with diabetes mellitus (DM). However, data on hepatocellular carcinoma (HCC) are very limited and inconsistent. This meta-analysis was designed to pool data currently available to determine the association between metformin use and HCC among diabetic patients. Methods. The Medline and Embase databases were searched to identify the relevant studies between January 1966 and December 2011. The overall analysis was derived using a random-effects meta-analysis model (DerSimonian and Laird method). Subgroup analysis was performed to explore the source of heterogeneity and validate the results from overall analysis. The Newcastle-Ottawa Quality assessment scales were adopted for quality assessment; Begg's funnel plot and Egger's regression asymmetry test were used to detect the publication bias. Results. A total of seven studies were identified, including three cohort studies and four case-control studies. Based on the available data, the overall prevalence of HCC was 3.40% (562/16,549) in DM patients. The overall analysis showed a significantly reduced risk of HCC in metformin users versus nonusers in diabetic patients (relative risk (RR) 0.24, 95% confidence interval (CI) 0.13-0.46, p < 0.001). Fifteen subgroup analyses were performed, and most of them (12/15 = 80%) provided supporting evidence for the results of overall analysis. Begg's (Z = -0.15, p = 0.8819) and Egger's test (t = -0.79, p = 0.468) showed no significant risk of having a publication bias. Conclusion. Metformin treatment was associated with reduced risk of HCC in diabetic patients. To clarify this relationship, more high-quality studies are required.