Background Little is known about the effectiveness of befriending for people with intellectual disability and whether or not befriending improves depressive symptoms and social outcomes. Objective This pilot trial aimed to assess the feasibility and acceptability of conducting a future full-scale randomised controlled trial of one-to-one befriending for people with intellectual disability who had depressive symptoms. Design This was a parallel-group, two-armed randomised controlled trial incorporating an exploratory economic analysis and a mixed-methods process evaluation. Outcome assessments were conducted at baseline and at 6 months post randomisation by a research assistant who was blind to allocation. We aimed to approach 50 participants, with a view to recruiting 40. Setting Participants with intellectual disability were recruited from one NHS trust and from referrals to two community befriending services. The intervention was delivered by community befriending services. Participants Adults with mild or moderate intellectual disability with a score on the Glasgow Depression Scale for people with a Learning Disability of ≥ 5 were included. Those attending a day service/college for ≥ 3 days a week were excluded. Volunteers were aged ≥ 18 years and had no history of prior convictions. Intervention Participants in the intervention group were matched with a volunteer befriender and were expected to meet once per week for 1 hour, over 6 months. Volunteers recorded activities in a logbook. Volunteers received training and regular supervision. Both groups received usual care and a resource booklet of local activities. Main outcome measures The feasibility outcomes and progression criteria were recruitment of at least 70% of participants approached; matching of at least 70% of participants in the intervention group to a volunteer; a dropout rate of < 30% of participants and volunteers; adherence to the intervention (10 meetings between pairs); acceptability of the intervention; and feasibility of collecting data on costs and resource use for an economic evaluation. Changes in depressive symptoms (primary clinical outcome: Glasgow Depression Scale) and self-esteem, quality of life, social participation, social support, health-related quality of life and service use were recorded at 6 months. Outcomes in volunteers were also assessed. Results We recruited only 16 participants with intellectual disability (40% of target) and 10 volunteers. Six of the eight (75%) participants in the intervention group were matched with a befriender and there was good adherence (mean number of meetings attended 11.8; range 1–21 meetings). Going to a cafe/restaurant and having a conversation were the most frequent activities. All participants were retained at follow-up, but two volunteers dropped out. Trial procedures and the intervention might be acceptable, but modifications were suggested. Data on costs and resource use were obtained, but there were discrepancies in the health-related quality-of-life data. Limitations Delays to the study prevented the use of alternative recruitment strategies and the planned 12-month follow-up could not be completed. Conclusions Recruitment was not feasible, but other feasibility outcomes were more positive. Future work Evaluating befriending for people with intellectual disability could be explored through alternative study designs, such as observational studies. Trial registration Current Controlled Trials ISRCTN63779614. Funding This project was funded by the National Institute for Health Research (NIHR) Public Health Research programme and will be published in full in Public Health Research; Vol. 9, No. 10. See the NIHR Journals Library website for further project information.