Background: At least two out of three persons with Parkinson’s disease (PD) experience falls each year. Many of these patients are recurrent fallers, primarily due to Parkinsonian symptoms that disrupt postural control and limit responses to balance challenges. Despite pharmacological and rehabilitation interventions, postural and balance deficits persist, and the risk of falling remains intractable. Among older adults, perturbation-based balance training (PBT)- where fall-inducing disturbances are applied unpredictably- has emerged as an effective method to reduce fall-related injuries. However, persons with PD exhibit set-shifting and postural inflexibility impairments that can hinder motor skill acquisition, retention, and generalization under unpredictable practice conditions. Consequently, motor learning strategies that are effective for those without PD may not yield the same results in this population. The overall objective of this dissertation was to investigate the effects of practice schedules on the motor learning of protective stepping stability in persons with PD and age-matched controls. We hypothesized that blocked practice would be more effective than random practice for acquiring and retaining protective stepping stability in response to unexpected balance perturbations. Furthermore, additional random practice will be necessary to generalize these protective skills to different contexts. Methods: We randomly assigned 20 persons with PD and 20 age-matched controls to a blocked and random practice group to receive a session of fall-inducing perturbations in blocked or random order while participants wore a safety harness. Additional random perturbations were administered to both groups 24 to 48 hours after the first practice session. We recorded their postural responses using motion capture systems that allowed us to quantify their postural responses in three dimensions and compared performance between practice groups at baseline, posttest, retention, and generalization tests. Result: Persons with PD retained protective stepping stability through blocked but not random practice after 24 to 48 hours of practice. With additional random practice, improvement persisted up to 48 hours to one week after practice. Conclusion: When designing PBT to improve reactive balance in people with PD, the choice motor learning impairments should be accounted for. Blocked practice may enhance the acquisition and retention of protective stepping through PBT in people with PD.