A number of studies have investigated the association between TNF-308 (rs1800629 G/A) polymorphisms and the susceptibility towards tuberculosis (TB) in different populations. However, many of these studies provided inconsistent results. In this study, a systematic review and meta-analysis of the published studies was performed to gain a clearer understanding of this association. The PubMed, Embase, Web of Science and CNKI databases were searched for case-control studies published up to Jan 2011, we used no lower date limit. Data were extracted and pooled odds ratios (OR) with 95% confidence intervals (CI) were calculated. A total of 18 publications from 2001 to 2010, involving 2584 TB cases and 3817 controls were included. Overall, for the A allele carriers (G/A + A/A) vs. homozygote GG, the pooled OR was 1.03 (95% CI = 0.89-1.19; P = 0.912 for heterogeneity). For the allele A vs. allele G, the pooled OR was 1.07 (95% CI = 0.93-1.22; P = 0.013 for heterogeneity). In the stratified analysis by ethnicity, among Asians significant risk was found for allele A vs. allele G (OR = 1.22, 95% CI = 1.02-1.47; P = 0.152 for heterogeneity), no significant risks were found among Caucasians. This meta-analysis indicated that the TNF-308 polymorphism was not associated with the risk of TB in the total population, however the significant risk for TNF-308 A allele was found among Asians not Caucasians.