Object: The primary aim of this study was to estimate the cutoff value of hemoglobin A1c (HbA1c) to identify diabetes in a high-risk population for type 2 diabetes. Methods: A total of 658 participants with a BMI of ≥ 23 kg/㎡ and who were aged ≥ 30 years but have an HbA1c value < 6.5% were selected from the community-based Korean Diabetes Prevention Study participants and enrolled into this study. Diabetes diagnosis and determination of increased diabetes risk (prediabetes) were carried out on the basis of the 2018 ADA criteria. The cutoff values of HbA1c for detecting diabetes and prediabetes were identified using the receiver operating characteristic (ROC) analysis. The gold standard for diabetes prediction was defined by the 2-h postprandial glucose (pp2hr) level of ≥ 200 mg/dL. Results: The participants comprised 54.3% women, and the mean age of all participants was 53.1 ± 7.6 years. At baseline, fasting plasma glucose was 98.1 ± 12.1 mg/dL, pp2hr value was 136.4 ± 43.3 mg/dL, and HbA1c was 5.7% ± 0.3%. Sixty-seven subjects (10.2%) were diagnosed with diabetes and 436 subjects (66.3%) were determined to be prediabetes. HbA1c provided the area under the ROC curve (AUC) for diabetes of 0.76, but AUC for impaired fasting glucose of 0.64, and impaired glucose tolerance of 0.56. The optimal HbA1c cutoff value for diagnosing diabetes was 6.0% (sensitivity 59.0%, specificity 80.0%) in this study. Conclusion: HbA1c cutoff value of 6.0% can be used as an optimal screening value for diabetes instead of the oral glucose tolerance test value, but it might be an inadequate screening tool for prediabetes in a high-risk population.