Introduction: Dyspnea is common and often severe symptom in mechanically ventilated patients. Proportional assist ventilation (PAV) is an assist ventilatory mode that adjusts the level of assistance to the activity of respiratory muscles. We hypothesized that PAV reduce dyspnea compared to pressure support ventilation (PSV). Patients and methods: Mechanically ventilated patients with clinically significant dyspnea were included. Dyspnea intensity was assessed by the Dyspnea-Visual Analog Scale (D-VAS) and the Intensive Care-Respiratory Distress Observation Scale (IC-RDOS) at inclusion (PSV-Baseline), after personalization of ventilator settings in order to minimize dyspnea (PSV-Personalization), and after switch to PAV. Respiratory drive was assessed by record of electromyographic activity of inspiratory muscles, the proportion of asynchrony was analyzed. Results: Thirty-four patients were included (73% males, median age of 66 [57-77] years). The D-VAS score was lower with PSV-Personalization (37 mm [20‒55]) and PAV (31 mm [14‒45]) than with PSV-Baseline (62 mm [28‒76]) (p < 0.05). The IC-RDOS score was lower with PAV (4.2 [2.4‒4.7]) and PSV-Personalization (4.4 [2.4‒4.9]) than with PSV-Baseline (4.8 [4.1‒6.5]) (p < 0.05). The electromyographic activity of parasternal intercostal muscles was lower with PAV and PSV-Personalization than with PSV-Baseline. The asynchrony index was lower with PAV (0% [0‒0.55]) than with PSV-Baseline and PSV-Personalization (0.68% [0‒2.28] and 0.60% [0.31‒1.41], respectively) (p < 0.05). Conclusion: In mechanically ventilated patients exhibiting clinically significant dyspnea with PSV, personalization of PSV settings and PAV results in not different decreased dyspnea and activity of muscles to a similar degree, even though PAV was able to reduce asynchrony more effectively.