Purpose: Exercise ECG has been known as a useful tool for evaluation of coronary artery disease (CAD) with 70% in sensitivity and specificity. However, false negative result in Treadmill test is still problematic to physician due to age, gender, medication, and medical history. The aim of this study was to determine whether the heart rate recovery (HRR) discriminate false negative from true negative result in exercise ECG. Methods: We enrolled 543 patients to visit to our hospital due to typical chest pain. Patients underwent symptom-limited exercise treadmill testing using modified Bruce protocols and Computer tomography (CT) or coronary angiography (CAG) for evaluation of CAD. We calculated the Duke treadmill exercise score (DTES) and divided the patients into low-, intermediate- and high-risk according to TES. Also, HRR was calculated as maximal HR minus recovery 1 min HR. We divided the patients with low risk DTES into two groups according to presence of significant CAD in CT or CAG ; true negative(TN) group (292 patients, 57.8짹10.0 yrs) and false negative(FN) group (251 patients, 55.3짹10.9 yrs). We compared age, gender, HRR, medications, medical history between both groups. Results: There was higher incidence of male and older in FN group. The HRR in FN group was significantly higher compared to it in TN group. In multivariate logistic analysis, the HRR was independent predictor for discrimination FN from TN group after adjustment of age, diabetes, hypertension and medication history like beta blocker and calcium channel blocker, and nitrate (p=0.046, odd ratio 0.986). Conclusions: The HRR might be a useful predictor for discrimination false negative from true negative result in exercise ECG. However, further larger study might be needed.