Background: The 2518 A/G polymorphism in the MCP-1 gene has been extensively studied for association swith cancer; however, results from replication studies have been inconsistent. The aim of this investigation was to determine links with risk of cancer by meta-analysis. Methods: We searched Pubmed, Embase, CNKI, Weipu and Wanfang databases, covering all case-control studies until March, 2013. Statistical analyses were performed using the Revman 5.0 software. Results: A total of 11 case-control studies met our inclusion criteria, including 1,422 cases and 2,237 controls. The results indicated that the MCP-1 2518 gene polymorphism had no association with cancer risk overall (GG vs.GA+ AA: OR = 0.89, 95%CI = 0.61-1.28, P = 0.52). However, in the subgroup analysis by ethnicity, a decrease of cancer risk was found in Asian populations (GG vs.GA+ AA: OR = 0.79, 95%CI = 0.63-0.99, P = 0.04). Conclusion: This meta-analysis suggested that the 2518A/G polymorphism of MCP-1 gene is associated with risk of cancer among Asian, but not in Caucasian populations.