Objective: To compare the effects of an inspiratory versus and expiratory muscle-training program on voluntary and reflex peak cough flow in patients with Parkinson disease. Design: A randomized controlled study. Setting: Home-based training program. Participants: In all, 40 participants with diagnosis of Parkinson’s disease were initially recruited in the study and randomly allocated to three study groups. Of them, 31 participants completed the study protocol (control group, n = 10; inspiratory training group, n = 11; and expiratory training group, n = 10) Intervention: The inspiratory and expiratory group performed a home-based inspiratory and expiratory muscle-training program, respectively (five sets of five repetitions). Both groups trained six times a week for two months using a progressively increased resistance. The control group performed expiratory muscle training using the same protocol and a fixed resistance. Main measures: Spirometric indices, maximum inspiratory pressure, maximum expiratory pressure, and peak cough flow during voluntary and reflex cough were assessed before and at two months after training. Results: The magnitude of increase in maximum expiratory pressure (d = 1.40) and voluntary peak cough flow (d = 0.89) was greater for the expiratory muscle-training group in comparison to the control group. Reflex peak cough flow had a moderate effect (d = 0.27) in the expiratory group in comparison to the control group. Slow vital capacity (d = 0.13) and forced vital capacity (d = 0.02) had trivial effects in the expiratory versus the control group. Conclusions: Two months of expiratory muscle-training program was more beneficial than inspiratory muscle-training program for improving maximum expiratory pressure and voluntary peak cough flow in patients with Parkinson’s disease.