Introduction: Fall rates in people with Parkinson's Disease range between 35 and 68% with the majority of falls occurring while walking. Initial evidence suggests that when walking without arm swing, people with Parkinson's Disease adapt their stepping foot placement as a means to preserve dynamic stability. However, it remains unexamined what arm swing's effect has on dynamic stability when walking on destabilizing surfaces. Methods: Twenty people with Parkinson's Disease (63.78 ± 8.97 years) walked with restricted and unrestricted arm swing on unperturbed, rocky, rolling-hills, and mediolateral translational surfaces. Data were collected on a split-belt treadmill CAREN Extended-System (Motek Medical, Amsterdam, NL). Bilateral averages and coefficient of variations for step time, length, and width; and mediolateral margin of stability were calculated. Results: Results were examined in three separate analyses that included arm conditions during each of the destabilizing surfaces compared to unperturbed walking (arm-rolling hills, arm-rocky, and arm-mediolateral). Compared to unrestricted arm swing, restricted arm swing reduced average step length (arm-rolling hills) and time (arm-rocky), and increased COV step time (arm-rolling hills). The arm-rolling hills analysis revealed that the most affected leg had a shorter step length than the least affected. The destabilizing surface effects revealed that during the arm-rolling hills and arm-rocky analyses, step time decreased, step width increased, and the COV for step time, length and width increased. No main effects occurred for the arm-mediolateral analysis. Conclusion: Results indicate that foot placement in response to restricted arm swing, in people with Parkinson's Disease, depends on the encountered destabilizing surface. The arm-rolling hills analysis revealed that participants appropriately reduced step length as compensation to their restricted arm swing. However, the arm-rocky analysis revealed that individuals prioritized forward progression over dynamic stability as they decreased average step time. Additionally, the increased spatiotemporal variability in response to the rocky and rolling hills conditions indicate partial foot placement adaptation to maintain an already existing level of global dynamic stability as no changes in the Margin of Stability occurred. Adaptation is further corroborated by the decreased step time and increased step width. These responses reflect attempts to pass the destabilizing terrains faster while increasing their base of support.