Background: The causes of exertional desaturation in patients with COPD can be multifactorial. We aimed to investigate factors predict exertional desaturation in patients with moderate to severe COPD. Methods: We tested 51 consecutive patients with stable COPD (FEV1/FVC, $40{pm}13%$ predicted). Patients performed a six minute walk test (6MWT). Pulse oxymetric saturation (SpO2) and pulse rate were recorded. Results: Oxygen desaturation was found in 15 subjects after 6MWT, while 36 subjects were not desaturated. Lung diffusing capacity was significantly lower in desaturation (DS) group ($62{pm}18%$ predicted) compared with not desaturated (ND) group ($84{pm}20$, p<0.01). However there was no statistical difference of FEV1/FVC ratio or residual volume between two groups. The pulse rate change was significantly higher in the desaturated compared with the not desaturated group. Six minute walking distance, subjective dyspnea scale, airflow obstruction, and residual volume did not predict exertional oxygen desaturation. Independent factors assessed by multiple logistic regression revealed that a pulse rate increment (odd ratio [OR], 1.19; 95% confidence interval [CI], 1.01~1.40; p=0.02), a decrease in baseline PaO2 (OR, 1.105; 95% CI, 1.003~1.218; p=0.04) and a decrease in lung diffusing capacity (OR, 1.10; 95% CI, 1.01~1.19; p=0.01) were significantly associated with oxygen desaturation. Receiver operator characteristic (ROC) analysis showed that an absolute increment in pulse rate of 16/min gave optimal discrimination between desaturated and not desaturated patients after 6MWT. Conclusion: Pulse rate increment and diffusion capacity can predict exertional oxygen desaturation in stable COPD patients with moderate to severe airflow obstruction.