We attempted to systematically elucidate the association between monocyte chemoattractant protein-1 (MCP-1) -2518A>G polymorphism and risk of coronary artery disease (CAD). Eligible studies were identified through PubMed, EBSCO, and Web of Science Databases. The magnitude of MCP-1 polymorphism effect and its possible mode of action on CAD were estimated. The odds ratio (OR) with 95% confidence intervals (CI) were pooled in a specific genetic model to assess the association. A total of 21 studies were involved. There was significant gene effect on CAD risk in the overall population (likelihood ratio test: p < 0.0001). Patients with GG and AG genotypes had 1.435 (95% CI: 1.183–1.740) and 1.087 (95% CI: 1.008–1.172) times higher risk of CAD than those with AA genotype. These gene effects suggested a recessive model to be appropriate. The pooled OR was 1.362 (95% CI: 1.137–1.631; puncorrected = 0.001, pFDR = 0.005) in the recessive model. In the ethnicity-stratified analysis, significant association was observed in the Caucasian population (OR = 1.492; 95% CI: 1.106–2.014; puncorrected = 0.009, pFDR = 0.015), whereas no statistical significant association was detected in the Asian population (adjusted p = 0.124). The results suggested that MCP-1 -2518A>G polymorphism may be associated with susceptibility to CAD, especially in Caucasians.