Abstract Background Many people with dementia experience anxiety, which can lead to decreased independence, relationship difficulties and increased admittance to care homes. Anxiety is often treated with antipsychotic medication, which has limited efficacy and serious side effects. Cognitive behavioural therapy (CBT) is widely used to treat anxiety in a range of populations, yet no RCTs on CBT for anxiety in dementia exist. This study aims to develop a CBT for anxiety in dementia manual and to determine its feasibility in a pilot RCT. Methods/design Phase I involves the development of a CBT for anxiety in dementia manual, through a process of (1) focus groups, (2) comprehensive literature reviews, (3) expert consultation, (4) a consensus conference and (5) field testing. Phase II involves the evaluation of the manual with 50 participants with mild to moderate dementia and anxiety (and their carers) in a pilot, two-armed RCT. Participants will receive either ten sessions of CBT or treatment as usual. Primary outcome measures are anxiety and costs. Secondary outcome measures are participant quality of life, behavioural disturbance, cognition, depression, mood and perceived relationship with the carer, and carer mood and perceived relationship with the person with dementia. Measures will be administered at baseline, 15 weeks and 6 months. Approximately 12 qualitative interviews will be used to gather service-users' perspectives on the intervention. Discussion This study aims to determine the feasibility of CBT for people with anxiety and dementia and provide data on the effect size of the intervention in order to conduct a power analysis for a definitive RCT. The manual will be revised according to qualitative and quantitative findings. Its publication will enable its availability throughout the NHS and beyond. Trial registration ISRCTN64806852