Cyclooxygenase-2 (COX-2) is proven to influence the carcinogenesis through immune response suppression, apoptosis inhibition, angiogenesis regulation, and tumor cell invasion. Previous studies assessing the association between COX-2 1195 G/A polymorphism and susceptibility to hepatocellular carcinoma (HCC) reported conflicting results. The objective of the study was to investigate the association between COX-2 1195 G/A polymorphism and HCC by a meta-analysis. PubMed, Embase, Web of Science, and Wangfang databases were searched for studies investigating the association between COX-2 1195 G/A polymorphism and HCC risk. The pooled odds ratio (OR) and its 95 % confidence interval (CI) were used to assess the strength of the association. Five studies with a total of 1,690 HCC cases and 1,961 controls were identified. Meta-analyses of total included studies showed that there was an obvious association between COX-2 1195 G/A polymorphism and HCC risk under two main genetic models (for AA versus GG, fixed-effects OR=1.45, 95 % CI 1.15-1.81, P=0.001, I (2) =0.0 %; for AA/GA versus GG, fixed-effects OR=1.26, 95 % CI 1.05-1.51, P=0.011, I (2) =0.0 %). Subgroup analysis by ethnicity showed that association was still obvious in Asians under two genetic models (for AA versus GG, fixed-effects OR=1.45, 95 % CI 1.16-1.82, P=0.001, I (2) =21.7 %; for AA/GA versus GG, fixed-effects OR=1.27, 95 % CI 1.05-1.54, P=0.013, I (2) =0.4 %). The evidence from the meta-analysis supports an association between COX-2 1195 G/A polymorphism and HCC risk in Asians. Further studies with large sample and careful design are needed to identify the possible association in Caucasians.