OBJECTIVES: Medical disease-modifying treatments are lacking in Parkinson's disease (PD). Physical exercise is potentially an inexpensive and easily accessible disease-modifying therapy, but this remains to be studied using recent promising magnetic resonance imaging (MRI) markers of disease progression. The objective was to determine whether 24 weeks of moderate to high-intensity progressive aerobic exercise affects brain MRI measures in people with PD (pwPD). METHODS: A 24-week randomized controlled multicenter trial (with a 24-week follow-up) was conducted, including an exercise group (supervised aerobic exercise followed by self-guided physical activity) and a control group (habitual lifestyle). Patients older than 40 years with mild to moderate impairments (Hoehn & Yahr score ≤3) were randomized (1:1). The primary outcome was R2* (effective transverse relaxation rate) after 24 weeks, which was analyzed using the intention-to-treat principle. Additional outcomes included free water and neuromelanin levels in the substantia nigra, cardiorespiratory fitness, the Movement Disorder Society-sponsored Revision of the Unified Parkinson's Disease Rating Scale (MDS-UPDRS), and levodopa equivalent daily dose (LEDD). RESULTS: Seventy participants were randomized to exercise (n = 36) or control (n = 34). R2* did not change over the intervention period (between-group difference: -1.14 [-5.14; 2.87], mean [95%CI]). Similarly, no significant effects were seen in cardiorespiratory fitness (+1.64 mL O2/min/kg [-0.23; 3.51]) and MDS-UPDRS III (-0.7 [-5.3; 4.0]), whereas LEDD increased less following exercise (-78.4 [-154.6; -2.1]). CONCLUSION: These findings do not support a neuroprotective effect of progressive aerobic exercise on R2* in pwPD, whereas the intervention appears to benefit the efficacy of levodopa. CLINICAL TRIAL REGISTRATION: Clinicaltrials.gov identifier: NCT04379778.