Genome-wide association studies have identified 6p21 as a new lung cancer susceptibility locus in populations of European descent. Since then, the relationship between common variations on 6p21 and lung cancer has been reported in various ethnic groups; however, these studies have yielded inconsistent results. To investigate this inconsistency, we performed a meta-analysis of 11 studies involving a total of 36,890 cases and 52,767 controls for three widely studies polymorphisms (rs4324798, rs3117582, and rs9295740) to evaluate its effect on genetic susceptibility for lung cancer. An overall random-effects per-allele odds ratio (OR) of 1.11 (95% confidence interval (CI) 1.04–1.19; P = 0.002) and 1.20 (95% CI 1.14–1.26; P < 10(−5)) was found for the rs4324798 and rs3117582 polymorphism, respectively. Marginal significant associations were also detected for rs9295740 with per-allele OR of 1.09 (95% CI 1.01–1.18; P = 0.03). In the subgroup analysis by ethnicity, significantly increased risks were found for the three polymorphisms among Caucasians. Similar results were also observed using dominant or recessive genetic models. This meta-analysis demonstrated that the three common variations (rs4324798, rs3117582, and rs9295740) on 6p21 are risk factors associated with increased lung cancer susceptibility, but these associations vary in different ethnic populations.