To describe changes in balance, walking speed, functional mobility, and eye movements following an activity-oriented physiotherapy (AOPT) or its combination with eye movement training (AOPT-E) in patients with Parkinson's disease (PD). To explore the feasibility of a full-scale randomised controlled trial (RCT). Using an assessor-blinded pilot RCT, 25 patients with PD were allocated to either AOPT or AOPT-E. Supervised interventions were performed 30 minutes, 4x/weekly, for 4 weeks, alongside inpatient rehabilitation. Outcomes were assessed at baseline and post-intervention, including dynamic balance, walking speed, functional and dual-task mobility, ability to safely balance, health-related quality of life (HRQoL), depression, and eye movements (number/duration of fixations) using a mobile eye tracker. Freezing of gait (FOG), and falls-related self-efficacy were assessed at baseline, post-intervention, and 4-week follow-up. Effect sizes of 0.10 were considered weak, 0.30 moderate, and ≥0.50 strong. Feasibility was assessed using predefined criteria: recruitment, retention and adherence rates, adverse events, falls, and post-intervention acceptability using qualitative interviews. Improvements were observed in dynamic balance (effect size r = 0.216-0.427), walking speed (r = 0.165), functional and dual-task mobility (r = 0.306-0.413), ability to safely balance (r = 0.247), HRQoL (r = 0.024-0.650), and depression (r = 0.403). Falls-related self-efficacy (r = 0.621) and FOG (r = 0.248) showed varied improvements, partly sustained at follow-up. Eye movement improvements were observed after AOPT-E only. Feasibility analysis revealed that recruitment was below target, with less than two patients recruited per month due to COVID-19 restrictions. Feasibility targets were met, with a retention rate of 96% (95% confidence interval [CI]: 77.68-99.79) and a 98.18% (95% CI: 96.12-99.20) adherence rate, exceeding the targets of 80% and 75%, respectively. One adverse event unrelated to the study intervention confirmed intervention safety, and interview data indicated high intervention acceptability. AOPT-E and AOPT appeared to be effective in patients with PD. Feasibility of a larger RCT was confirmed and is needed to validate results.