Background: Non-ambulatory people with multiple sclerosis comprise 25% of the multiple sclerosis population. Literature reviews show insufficient evidence exists regarding physiotherapy for this population. A qualitative study suggested benefits from ¿exercise buddies¿, who were paid carers delivering a physiotherapy home exercise programme. Aims: To explore the feasibility and effects of ¿exercise buddies¿ for non-ambulatory people with multiple sclerosis. Methods: Twenty-nine non-ambulatory people with multiple sclerosis (age range: 43¿72) were randomised to 10 weeks of ¿usual care¿ or the ¿exercise buddy¿ intervention. People with multiple sclerosis were assessed with the Multiple Sclerosis Impact Scale 29 and the Guys Neurological Disability Scale pre- and post-intervention. Their informal caregivers (12 male, 16 female, aged 21¿68) completed the Adult Carer Quality of Life questionnaire. Findings: Using analysis of covariance to adjust for pre-intervention scores, there was no significant differences between groups after treatment on the Multiple Sclerosis Impact Scale 29 physical (P=0.395), Multiple Sclerosis Impact Scale 29 psychological (P=0.176) or Guys Neurological Disability Scale (P=0.177). The Adult Carer Quality of Life was also not significantly different between groups post-treatment (P=0.432). Using paired t-tests, the exercise buddy group improved significantly from baseline on the two components of the Multiple Sclerosis Impact Scale 29 (physical: P=0.024; psychological: P=0.009), which was not seen in the usual care group. Conclusions: This pilot study found no significant between group differences post-treatment. However, good feasibility and significant positive changes from baseline for the exercise buddy group warrant further exploratory work, in addition to a cost analysis.