Background Many people with dementia living in care homes have distressing and costly agitation symptoms. Interventions should be efficacious, scalable, and feasible. Methods We did a parallel-group, cluster-randomised controlled trial in 20 care homes across England. Care homes were eligible if they had 17 residents or more with dementia, agreed to mandatory training for all eligible staff and the implementation of plans, and more than 60% of eligible staff agreed to participate. Staff were eligible if they worked during the day providing face-to-face care for residents with dementia. Residents were eligible if they had a known dementia diagnosis or scored positive on screening with the Noticeable Problems Checklist. A statistician independent of the study randomised care homes (1:1) to the Managing Agitation and Raising Quality of Life (MARQUE) intervention or treatment as usual (TAU) using computer-generated randomisation in blocks of two, stratified by type of home (residential or nursing). Care home staff were not masked to the intervention but were asked not to inform assessors. Residents with dementia, family carers, outcome assessors, statisticians, and health economists were masked to allocation until the data were analysed. MARQUE is an evidence-based manualised intervention, delivered by supervised graduate psychologists to staff in six interactive sessions. The primary outcome was agitation score at 8 months, measured using the Cohen-Mansfield Agitation Inventory (CMAI). Analysis of the primary outcome was done in the modified intention-to-treat population, which included all randomly assigned residents for whom CMAI data was available at 8 months. Mortality was assessed in all randomly assigned residents. This study is registered with the ISRCTN registry, number ISRCTN96745365.