Background: Age is the most important risk factor for development of dementia and the recommendation is that the elderly be cognitively tested in order to detect impairment in the initial phase for adequate treatment. The demand for the care of these elderly people is great, drawing attention to the need for rapid tests, with good accuracy and simple application to identify cognitive impairment. Objective: To develop the M-ACE Brazilian version using data from ACE-R deriving sub-items that could better predict the diagnosis of cognitive impairment. Methods: The M-ACE BR was developed using Mokken scaling analysis in 352 participants (cognitively normal = 232, cognitive impairment no dementia (CIND) = 82 and dementia = 38) and validated in an independent sample of 117 participants (cognitively normal = 25, CIND = 88 and dementia = 4). Results: The M-ACE BR has nine items (spatial orientation, anterograde memory, retrograde memory, delayed recall, recognition, verbal fluency letter “P”, repetition of four words, naming 10 items and comprehension) with a max. score of 51 points and average duration time of seven minutes. The cutoff score ≤43/51 for CIND had a sensitivity of 59.09% and a specificity of 80%. For a screening test in which sensitivity is prioritized for further investigation, we suggest using a cutoff of ≤47 (sensitivity 85.23% and specificity 24%), maintaining a good positive predictive value (79.8%) Conclusion: The M-ACE BR is a brief and adequate instrument for detecting cognitive impairment in elderly Brazilians.