Background Looked-after children are at risk of suboptimal attachment patterns and reactive attachment disorder. However, access to interventions varies widely and there are no evidence-based interventions for this disorder. Objectives (1) To adapt an existing video-feedback intervention to meet the specific needs of foster children in the UK with reactive attachment symptoms, (2) to conduct a case series to road-test the treatment manual and study procedures, (3) to conduct a scoping study of the key hurdles in a pilot trial and (4) to conduct a pilot randomised controlled trial of the adapted intervention to determine the feasibility of a future full-scale trial. Design This was a mixed-methods study. The adapted treatment manual was developed with expert input and tested on a small case series. Qualitative interviews with key stakeholders were used in the scoping study in preparation for the trial and later with foster carers who received the new intervention. The final stage was a feasibility and pilot randomised controlled trial of the new intervention, compared with usual care. Researchers assessing the outcomes were blinded to group assignment. Setting The study was set in outpatient child and adolescent mental health services and partner social services departments. Sites included urban and rural/semirural areas. Participants Participants were foster carers with children aged ≤ 6 years presenting with difficulties in the domain of reactive attachment disorder. Key stakeholders included children’s services managers and mental health service practitioners in the scoping study. Foster carers who received the modified intervention participated in qualitative interviews. Intervention The video-feedback intervention to promote positive parenting and sensitive discipline is an extensively evaluated and effective treatment approach. This intervention was modified (based on the adapted version for foster care in the Netherlands) to suit the needs of young children with reactive attachment symptoms in foster care in the UK and was delivered to improve the sensitive responding of foster carers, foster carer–child relationships and child outcomes. The modified intervention was delivered in-home by trained mental health professionals over a period of 4–6 months. Main outcome measure The main outcome was reactive attachment symptom scores on the Disturbances of Attachment Interview. Results A series of minor changes to the intervention programme were introduced, which focused on improving its suitability for the UK foster care context. Challenges in recruitment meant that, despite numerous modifications to the protocol and the inclusion of additional sites, only 30 families (target, n = 40) were recruited to the randomised controlled trial (15 allocated to each group). However, most other trial parameters were deemed feasible and acceptable, particularly the high levels of data and treatment completeness. All randomised families were available for baseline analyses, but two in the treatment arm were not available for post-treatment analyses. The revised intervention was positively received by practitioners and foster carers. Limitations Only three-quarters of the target sample size was recruited. Furthermore, the sites’ own exclusion of potential participants and the low return rates of screening questionnaires raise the possibility of non-randomness of non-responses. Conclusion A larger-scale trial may be feasible, but only if recruitment barriers can be overcome. Dedicated resources to support recruitment within local authorities and wider inclusion criteria are recommended. Central resourcing of intervention capacity to supplement NHS staff is also recommended. Trial registration This trial is registered as ISRCTN18374094. Funding This project was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 26, No. 35. See the NIHR Journals Library website for further project information.