Signal averaged electrocardiogram (SAECG) is considered an important noninvasive indicator identifying patients at risk for ventricular arrhythmias. The aim of this study was to improve noninvasive prediction of CAD by integrating SAECG with the result of exercise tests in diabetic patients. Diabetic patients with stable angina pectoris underwent exercise testing and SAECG. Then a diagnostic score was derived that combined results of exercise testing and SAECG. A diagnostic score (0 to 2 points) was calculated by assigning 1 point for a positive exercise test result and 1 point for a positive SAECG. One hundred and seventy patients were included in the study. In patients with a score of 0, the likelihood of CAD is 18% whereas the likelihood of CAD is 95% in patients with a score of 2. Triple vessel CAD is present in 54%, 7% and only 1.5% of patients with score 2, 1 and 0 respectively. Therefore, patients with score 2 have a poor prognosis compared with score 1 or 0. A diagnostic score combining exercise testing and SAECG can distinguish patients with CAD from those without CAD with high accuracy in diabetic patients.