Type 2 diabetes mellitus (T2DM) is a global health problem that results from the interaction of environmental factors with genetic variants. Although a number of studies have suggested that genetic polymorphisms in the fat mass and obesity-associated () gene are associated with T2DM risk, the results have been inconsistent. To investigate whether polymorphisms associate with T2DM risk and whether this association is region-related, we performed this spatial analysis and meta-analysis. More than 60,000 T2DM patients and 90,000 controls from 62 case-control studies were included in this study. Odds ratios (ORs), 95% confidence intervals (CIs) and Moran's I statistic were used to estimate the association between rs9939609, rs8050136, rs1421085, and rs17817499, and T2DM risk in different regions. rs9939609 (OR = 1.15, 95% CI 1.11-1.19) and rs8050136 (OR = 1.14, 95% CI 1.10-1.18) conferred a predisposition to T2DM. After adjustment for body mass index (BMI), the association remained statistically significant for rs9939609 (OR = 1.11, 95% CI 1.05-1.17) and rs8050136 (OR = 1.08, 95% CI 1.03-1.12). In the subgroup analysis of rs9939609 and rs8050136, similar results were observed in East Asia, while no association was found in North America. In South Asia, an association for rs9939609 was revealed but not for rs8050136. In addition, no relationship was found with rs1421085 or rs17817499 regardless of adjustment for BMI. Moran's I statistic showed that significant positive spatial autocorrelations existed in rs9939609 and rs8050136. Studies on rs9939609 and rs8050136 focused on East Asia and South Asia, whereas studies on rs1421085 and rs17817499 were distributed in North America and North Africa. Our data suggest that the associations between rs9939609, rs8050136 and T2DM are region-related, and the two single-nucleotide polymorphisms contribute to an increased risk of T2DM. Future studies should investigate this issue in more regions.