Study Protocol ; BACKGROUND: Up to 90 % of people living with dementia in care homes experience one or more behaviours that staff may describe as challenging to support (BSC). Of these agitation is the most common and difficult to manage. The presence of agitation is associated with fewer visits from relatives, poorer quality of life and social isolation. It is recommended that agitation is treated through psychosocial interventions. Dementia Care Mapping™ (DCM™) is an established, widely used observational tool and practice development cycle, for ensuring a systematic approach to providing person-centred care. There is a body of practice-based literature and experience to suggests that DCM™ is potentially effective but limited robust evidence for its effectiveness, and no examination of its cost-effectiveness, as a UK health care intervention. Therefore, a definitive randomised controlled trial (RCT) of DCM™ in the UK is urgently needed. METHODS/DESIGN: A pragmatic, multi-centre, cluster-randomised controlled trial of Dementia Care Mapping (DCM™) plus Usual Care (UC) versus UC alone, where UC is the normal care delivered within the care home following a minimum level of dementia awareness training. The trial will take place in residential, nursing and dementia-specialist care homes across West Yorkshire, Oxfordshire and London, with residents with dementia. A random sample of 50 care homes will be selected within which a minimum of 750 residents will be registered. Care homes will be randomised in an allocation ratio of 3:2 to receive either intervention or control. Outcome measures will be obtained at 6 and 16 months following randomisation. The primary outcome is agitation as measured by the Cohen-Mansfield Agitation Inventory, at 16 months post randomisation. Key secondary outcomes are other BSC and quality of life. There will be an integral cost-effectiveness analysis and a process evaluation. DISCUSSION: The protocol was refined following a pilot of trial procedures. Changes include replacement of a questionnaire, whose wording caused some residents distress, to an adapted version specifically designed for use in care homes, a change to the randomisation stratification factors, adaption in how the staff measures are collected to encourage greater compliance, and additional reminders to intervention homes of when mapping cycles are due, via text message. TRIAL REGISTRATION: Current Controlled Trials ISRCTN82288852 . Registered on 16 January 2014. Full protocol version and date: v7.1: 18 December 2015. ; This project was funded by the National Institute for Health Research Health Technology Assessment programme (project number 11/15/13).