Physical activity is important for people with Parkinson's disease (PD) to improve disease-specific impairment and ameliorate secondary consequences related to deconditioning. Activity may also have a neuroprotective role if instigated early. Ambulatory activity has not been examined in incident PD. Eighty-nine newly diagnosed PD cases [mean (SD) age 67.3 (9.9) years] and 97 controls [mean (SD) 69.2 (7.7) years] wore an activity monitor (activPAL™) for 7 days. Volume, pattern and variability outcomes were compared. Accumulation of activity (α) was classified as short (< 30 s), medium (30 s-2 min) and long (> 2 min) bouts of walking. Associations between sustained walking (> 2 min) and motor, cognitive and affective characteristics were identified. Activity outcomes were considered with respect to global health recommendations. Total steps (volume), accumulation of bout length (α), and variability (S2w) outcomes were significantly different (all P < 0.001). PD participants (including Hoehn & Yahr (H&Y) stage I) accumulated significantly less time in long bouts (> 2 min) of walking compared with controls, due to performing fewer long bouts, rather than a reduction in time spent in walking per bout. For PD and controls there were weak but significant correlations for a range of characteristics and sustained walking. Fewer people with PD achieved the recommended 30 min of walking per day comprised of bouts > 10 min (P = 0.02) and bouts > 2 min (P < 0.001). People with PD were significantly less active than controls, with an inability to sustain levels of walking, and with differences apparent very early on in the disease process. A focus on increasing general ambulatory activity and exercise from the outset is recommended.