Balance training has been demonstrated to improve postural control in patients with Parkinson's disease (PD). The objective of this pilot randomized clinical trial was to investigate whether a balance training program using augmented visual feedback is feasible, safe, and more effective than conventional balance training in improving postural control in patients with PD. Methods: Thirty-three patients with idiopathic PD participated in a five-week training program consisting of ten group treatment sessions of 60min. Participants were randomly allocated to (1) an experimental group who trained on workstations consisting of interactive balance games with explicit augmented visual feedback (VFT), or (2) a control group receiving conventional training. Standing balance, gait, and health status were assessed at entry, at six weeks, and at twelve weeks follow-up. Results: Sixteen patients were allocated to the control group and seventeen to the experimental group. The program was feasible to apply and took place without adverse events. Change scores for all balance measures favored VFT, but the change in the primary outcome measure, i.e. the Functional Reach test, did not differ between groups (t(28)=-0.116, p =908). No other differences between groups were statistically significant. Conclusions: VFT proved to be a feasible and safe approach to balance therapy for patients with PD. In this proof-of-concept study VFT was not superior over conventional balance training although observed trends mostly favored VFT. These trends approached clinical relevance only in few cases: increasing the training load and further optimization of VFT may strengthen this effect. Trial registration: Controlled Trials, ISRCTN47046299.