Objective : Cardiopulmonary exercise test is a useful tool to evaluate the operative risk and to plan exercise treatment for the patients with chronic obstructive pulmonary disease(COPD). In cardiopulmonary exercise test, most of the measured parameters are recorded at the time of peak exercise, which are hard to attain in COPD patients. So we evaluated the usefulness of the parameter, breathing reserve index(BRI=minute ventilation [$V_E$]/maximal voluntary ventilation[MVV]) at the time of anaerobic threshold($BRI_{AT}$) for the differentiation of COPD patients with normal controls. Methods : Thirty-six COPD patients and forty-two healthy subjects underwent progressive, incremental exercise test with bicycle ergometer upto possible maximal exercise. All the parameters was measured by breath by breath method. Results : The maximal oxygen uptake in COPD patients (mean$pm$SE) was $1061.2{pm}65.6ml/min$ which was significantly lower than $2137.6{pm}91.4ml/min$ of normal subjects(p<0.01). Percent predicted maximal oxygen uptake was 54.3% in COPD patients and 86.0% in normal subjects(p<0.01). Maximal exercise(respiratory quotient; $VCO_2/VO_2{geq}1.09$) was accomplished in 7 of 36 COPD patients(19.4%) and in 18 of 42 normal subjects(42.9%). The $BRI_{AT}$ of COPD patients was higher($0.50{pm}0.03$) than that of control subject($028{pm}0.02$, p<0.01), reflecting early hyperventilation in COPD patient during exercise. The correlation between $BRI_{AT}$ and BRI at maximal exercise in COPD patients was good(r=0.9687, p<0.01). Conclusion : The $BRI_{AT}$ could be a useful parameter for the differentiation of COPD patients with normal controls in the submaximal cardiopulmonary exercise test.