Background. Parkinson’s disease (PD) causes difficulties with hand movements, which few studies have addressed therapeutically. Training with action observation (AO) and motor imagery (MI) improves performance in healthy individuals, particularly when the techniques are applied simultaneously (AO + MI). Both AO and MI have shown promising effects in people with PD, but previous studies have only used these separately. Objective. This article describes the development and pilot testing of an intervention combining AO + MI and physical practice to improve functional manual actions in people with PD. Methods. The home-based intervention, delivered using a tablet computer app, was iteratively designed by an interdisciplinary team, including people with PD, and further developed through focus groups and initial field testing. Preliminary data on feasibility were obtained via a six-week pilot randomised controlled trial (ISRCTN 11184024) of 10 participants with mild to moderate PD (6 intervention; 4 treatment as usual). Usage and adherence data were recorded during training, and semistructured interviews were conducted with participants. Exploratory outcome measures included dexterity and timed action performance. Results. Usage and qualitative data provided preliminary evidence of acceptability and usability. Exploratory outcomes also suggested that subjective and objective performance of manual actions should be tested in a larger trial. The importance of personalisation, choice, and motivation was highlighted, as well as the need to facilitate engagement in motor imagery. Conclusions. The results indicate that a larger RCT is warranted, and the findings also have broader relevance for the feasibility and development of AO + MI interventions for PD and other conditions.