In this paper, based on the association of methylenetetrahdrofolate reductase (MTHFR) with type 2 diabetes, we evaluated the association of polymorphism with morbidly obesity on risk type 2 diabetes. A case-control study of 74 health morbidly obese and 76 healthy non-obese was conducted in Iraq. MTHFR (C677T, A1298C and G1793A) genotyping was performed by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP). The study revealed a significant association between cytidine/thymine (CT) genotype of C677T and morbid obesity (odds ratio (OR): 1.26, 95% confidence interval (CI): 0.91-1.60, probability (P) = 0.0003), but did not show any significant association in thymine/thymine (TT) genotype (OR: 0.258, %95CI: 0.88-1.60, P = 0.082). On the other hand the second single nucleotide polymorphism (SNP) A1298C showed highly significant association in adenine/adenine (AA) genotype (OR: 1.39, %95CI: 0.91-1.58, P = 0.0001) and significant association between adenine/cytidine (AC) and cytidine/cytidine (CC) genotypes (OR: 0.702, %95CI: 0.88-1.61, P = 0.0377), (OR: 0.844, %95CI: 0.91-1.59, P = 0.0273), respectively, but showed no significant relation in three types of normal homozygous, heterozygous and rare homozygous in MTHFR-G1793A. The results suggested that A1298C substitution might pose a direct effect on being type 2 diabetes in morbidly obese patients, and C677T had moderate effect, while G1793A had no effect. However, further case-control studies are required to provide a more robust conclusion.