OBJECTIVE To investigate the risk of incident dementia according to fasting glucose levels and presence of comorbidities. RESEARCH DESIGN AND METHODS Using a health insurance claims database and the results of biennial health examinations in South Korea, we selected 8,400,950 subjects aged ≥40 years who underwent health examinations in 2009–2010. We followed them until 2016. Subjects’ baseline characteristics were categorized by presence of diabetes (yes/no) and glycemic status as normoglycemia, impaired fasting glucose (IFG), new-onset diabetes, or known diabetes (duration <5 years or ≥5 years). We estimated adjusted hazard ratios (aHRs) for dementia occurrence in each category. RESULTS During the observation period of 48,323,729 person-years, all-cause dementia developed in 353,392 subjects (4.2%). Compared with normoglycemia, aHRs (95% CI) were 1.01 (1.01–1.02) in IFG, 1.45 (1.44–1.47) in new-onset diabetes, 1.32 (1.30–1.33) in known diabetes <5 years, and 1.62 (1.60–1.64) in known diabetes ≥5 years. We found that associations between ischemic heart disease and chronic kidney disease with incident dementia were affected by the presence of diabetes. Ischemic stroke showed a greater association with incident dementia than diabetes. CONCLUSIONS Mild degrees of hyperglycemia and presence of comorbidities were associated with incident dementia. Intervention during the prodromal stage of a chronic disease (e.g., prediabetes) could be considered for dementia prevention.