Objectives: In order to investigate the effectiveness of pharmacological and non-pharmacological treatment methods for decreasing agitation in diseases like dementia, a proper scale like Cohen-Mansfield Agitation Inventory (CMAI) is needed. The aim of the present study was to evaluate the psychometric properties of the CMAI. Method: The sample consisted of 100 caregivers of patients with Alzheimer's disease that responded to the CMAI. Results: The inter-rater reliability of the inventory was investigated by the assessment of other observer. After two weeks, the main observer filled the CMAI for the same patients again. The reliability of the inventory was investigated by test-retest, alpha Cronbach and split-half methods that were 0.99, 0.92 and 0.82 respectively. The validity of the questionnaire was investigated by convergent validity, inter-rater agreement across items and exploratory factor analysis. The results from factor analysis of the CMAI using varimax rotation method yielded 4 factors labeled: 1- Aggressive Behaviors, 2- Non-Aggressive Behaviors, 3- Verbal Aggression, and 4- Hiding Behaviors, which explained 72.77% of the total variance. Convergent validity was confirmed by computing a correlation coefficient between the subscales together and with total scale, which was significant between 0.43 and 0.90 (p<0.01). Conclusion: Regarding the results, it could be said that this inventory can be a valid and reliable scale to evaluate the agitation in patients with Alzheimer's disease.