The purpose of this study was to investigate the effect of finger movement on ankle control for gait initiation in patients with Parkinson's disease (PD patients). The subjects were 13 PD patients and 6 age-matched healthy adults. The subjects moved fingers before or after gait initiation, or initiated gait without finger movement. Ankle joint movement in the stance leg was recorded to estimate the duration of ankle dorsiflexion (DIF duration), which reflects the degree of disturbance in ankle control for gait initiation in PD patients. In the PD patients with prolonged D/F duration, finger movement that preceded gait initiation shortened the D/F duration, but in the PD patients without prolonged D/F duration and in healthy subjects, the effect was not found. Accordingly, finger movement that precedes gait initiation improves ankle control for gait initiation in PD patients who suffer disturbance in ankle control for gait initiation.