Helium–oxygen mixtures and pressure-support ventilation have been used to unload the respiratory muscles and increase exercise tolerance in COPD. Considering the different characteristics of these techniques, we hypothesized that helium–oxygen would be more effective in reducing exercise-induced dynamic hyperinflation than pressure-support. We also hypothesized that patients would experience greater increases in respiratory rate and minute ventilation with helium–oxygen than with pressure-support. The hypotheses were tested in ten patients with severe COPD (FEV₁ = 28 ± 3% predicted [mean ± SE]) during constant-load cycling (80% maximal workrate) while breathing 30% oxygen-alone, helium–oxygen, and pressure-support in randomized order. As hypothesized, helium–oxygen had greater impact on dynamic hyperinflation than did pressure-support (end-exercise; p = 0.03). For the most part of exercise, respiratory rate and minute ventilation were greater with helium–oxygen than with pressure-support (p ≤ 0.008). During the initial phases of exercise, helium–oxygen caused less rib-cage muscle recruitment than did pressure-support (p < 0.03), and after the start of exercise it caused greater reduction in inspiratory reserve volume (p ≤ 0.02). Despite these different responses, helium–oxygen and pressure-support caused similar increases in exercise duration (oxygen-alone: 6.9 ± 0.8 min; helium–oxygen: 10.7 ± 1.4 min; pressure-support: 11.2 ± 1.6 min; p = 0.003) and similar decreases in inspiratory effort (esophageal pressure-time product), respiratory drive, pulmonary resistance, dyspnea and leg effort (p < 0.03). In conclusion, helium–oxygen reduced exercise-induced dynamic hyperinflation by improving the relationship between hyperinflation and minute ventilation. In contrast, pressure-support reduced hyperinflation solely as a result of lowering ventilation. Helium–oxygen was more effective in reducing exercise-induced dynamic hyperinflation in severe COPD, and was associated with greater increases in respiratory rate and minute ventilation than pressure-support.