Interstitial lung disease (ILD) is characterized by dyspnoea on exertion and exercise-induced hypoxaemia. High-flow nasal cannula (HFNC) therapy reduces the respiratory workload through higher gas flow and oxygen supplementation, which may affect exercise tolerance. This study aimed to examine the effects of oxygen and gas flow rates through HFNC therapy on exercise tolerance in ILD patients. We conducted three-treatment crossover study. All ILD patients performed the exercises on room air (ROOM AIR setting: flow, 0 L/min; fraction of inspired oxygen [FiO2], 0.21), HFNC (FLOW setting: flow 40 L/min, FiO2 0.21), and HFNC with oxygen supplementation (FLOW + OXYGEN setting: flow 40 L/min, FiO2 0.6). The primary endpoint was the endurance time, measured using constant-load cycle ergometry exercise testing at a peak work rate of 80%. Twenty-five participants (10 men, 71.2 ± 6.7 years) were enrolled. The increase in exercise duration between the ROOM AIR and FLOW was 46.3 s (95% CI, -6.1 to 98.7; p = 0.083), and the FLOW and FLOW + OXYGEN was 91.5 s (39.1-143.9; p < 0.001). The percutaneous oxygen saturation (SpO2) at rest was significantly higher with the FLOW + OXYGEN setting than with the ROOM AIR and FLOW settings, and the difference persisted during exercise. At equivalent time points during exercise, the SpO2 with the FLOW setting was significantly higher than that with the ROOM AIR setting. Oxygen supplementation in HFNC therapy improved exercise tolerance and SpO2. We found that gas flow alone did not improve exercise tolerance, but improved SpO2 during exercise.