Background: Immersive augmented reality (IAR) offers flexibility for designing rehabilitation tasks. The role of context in motor performance in persons with Parkinson disease (PD) suggests value in exploring the use of IAR in this population. Objective: Use IAR in a cohort of people with PD and controls to determine: 1) Does motor performance of PD differ between real-world (RW) and IAR tasks?; 2) Is this effect different in PwPD versus controls?; and 3) Does performance change with repeated exposure to the system? Method: Twenty-two participants with PD (15 male, mean age = 67.1 ± 6.1 years, mean disease duration = 6.2 ± 3.3 years) and 11 controls (6 male, mean age = 64.1 ± 5.6 years) completed functionally relevant motor tasks in two IAR scenarios: 1) Watering plants in a living room; and 2) Placing items in baskets at a grocery store. Comparable RW motor tasks were also performed. Visits were once per week for 3 weeks. Results: There was a significant effect of group (F(1, 31) = 4.980, p = 0.033, η2p = 0.138) and environment (Pillai's F(1, 31) = 98.519, p < 0.001, η2p = 0.761) in the living room task. For the grocery store scene, there was a multivariate effect of condition (Pillai's Trace F(2, 28) = 19.323, p < 0.05, η2 = 0.58), of visit (Pillai's Trace F(4, 116) = 3.371, p < 0.05, η2 = 0.208), and a within subjects multivariate condition x visit interaction (Pillai's Trace F(4, 116) = 4.487, p < 0.002, η2 = 0.265. Conclusions: An IAR program was developed to simulate two scenes and simple tasks were performed. Differences in performance between the IAR and RW environments may be due to altered delivery of visual information.