What specific attributes of exercise programs influence the preferences of people with Parkinson's disease for additional exercise compared with their current practice? What trade-offs are participants willing to make between exercise program attributes? Discrete choice experiment. Five hundred and forty people with Parkinson's disease. Participants decided whether they would adopt a hypothetical program in addition to their current exercise routine. Exercise program attributes included: type, number of sessions/week, location, travel time/session, delivery mode, supervisor's expertise, extent of supervision, benefits for physical and psychological function and out-of-pocket cost/session. Participants preferred additional exercise when programs: provided physical (OR 1.85, 95% CI 1.61 to 2.13) or psychological (OR 1.45, 95% CI 1.26 to 1.67) benefit, involved less travel time (ORs 1.50 to 2.02) and were supervised by qualified professionals with Parkinson's disease expertise (ORs 1.51 to 1.91). Participants were most willing to add multimodal exercise to their exercise routine (ORs 2.01 to 2.19). Participants were less likely to prefer higher cost programs (OR 0.65, 95% CI 0.60 to 0.71, per AU$10 cost increase) or group sessions compared to individual sessions (OR 0.72, 95% CI 0.54 to 0.96). Men preferred adding strengthening exercises (OR 2.00, 95% CI 1.23 to 3.26) and women had a preference against adding aerobic exercise (OR 0.33, 95% CI 0.15 to 0.73). Participants not currently exercising were more likely to prefer adding exercise compared with those already exercising 300 minutes weekly (OR 1.74, 95% CI 1.15 to 2.63). People with Parkinson's disease were more willing to participate in exercise programs that cost less, involve less travel, provide physical or psychological benefits and are supervised by qualified professionals. To enable more people with Parkinson's disease to exercise, health services should provide programs addressing these factors and account for sex differences.