Objective: To determine the effects of participation in a 2-year community-based dance class on disease severity and functional mobility in people with Parkinson disease (PD). Design: Randomized controlled trial. Settings/Location: Dance classes took place in a community-based location. Outcome measures were collected in a university laboratory. Patients: Ten individuals with PD were randomly assigned to the Argentine tango (AT) group (n = 5 [4 men]; mean age ± standard deviation, 69.6 ± 6.6 years) or the control group (n = 5 [4 men]; mean age ± standard deviation, 66 ± 11.0 years). Interventions: The AT group participated in a community-based AT class for 1 hour twice weekly for 2 years. Control group participants were given no prescribed exercise. Blinded assessments occurred at baseline and 12 and 24 months. Outcome measures: Movement Disorder Society-Unified Parkinson Disease Rating Scale (MDS-UPDRS) III, Mini-Balance Evaluation Systems Test (Mini-BESTest), gait velocity (forward and backward), Timed Up and Go and dual-task Timed Up and Go, Six-Minute Walk Test (6MWT), MDS-UPDRS II, MDS-UPDRS I, and Freezing of Gait Questionnaire. Results: There were no differences between groups at baseline. A significant group-by-time interaction (F [2,8] = 17.59; p < 0.0001) was noted for the MDS-UPDRS III, with the AT group having lower scores at 12 and 24 months than the controls. Significant interactions were also noted for the Mini-BESTest, MDS-UPDRS II and I, and 6MWT. Conclusion: This is believed to be one of the longest-duration studies to examine the effects of exercise on PD. Participation in community-based dance classes over 2 years was associated with improvements in motor and nonmotor symptom severity, performance on activities of daily living, and balance in a small group of people with PD. This is noteworthy given the progressive nature of PD and the fact that the control group declined on some outcome measures over 2 years.