Chronic ischemic heart disease (CIHD) is a frequent complication in patients with type 2 diabetes mellitus (T2DM), and its early diagnosis is crucial to improve the patient’s treatment and prognosis. However, diagnosis is complicated because symptoms are manifested unusually, and risk factors are diverse. In this research, a Transformer network model adapted for medical tabular data (TabTransformer), capable of diagnosing CIHD from clinical records of patients with T2DM, is developed, implemented, optimized, and evaluated. A dataset with longitudinal information from Mexican patients was used, achieving an accuracy of 87.72%, a recall of 90.16%, and a value of 0.9434 for the area under the ROC curve. The TabTransformer model significantly outperformed traditional machine learning approaches, demonstrating a better ability to capture complex relationships between clinical variables. The results confirm the effectiveness of the proposed approach for diagnosing CIHD in diabetic patients.