In the present study we analyzed the psychometric characteristics of the MoCA (Portuguese version) using the Rasch model for dichotomous items. The total sample comprised of 897 participants distributed between two main subgroups: (I) healthy group that was comprised of 650 cognitively healthy community dwellers and (II) clinical group that was comprised of 90 patients with Mild Cognitive Impairment, 90 patients with Alzheimer's disease, 33 patients with frontotemporal dementia, and 34 patients with vascular dementia recruited at a reference dementia clinic. All patients were investigated through a comprehensive neuropsychological assessment, laboratory tests essential to exclude a reversible form of dementia, imaging studies (CT or MRI and SPECT or FDG-PET), Apolipoprotein E allele genotyping and CSF biomarker (Aβ42,Tau, and P-tau) analyses. The clinical diagnosis was established through the consensus of a multidisciplinary team, based on international criteria. The results demonstrated an overall good fit of both items and the person's values, a high variability on cognitive performance level, and a good quality of the measurements. The MoCA scores also demonstrated adequate discriminant validity, with high diagnostic value. DIF analyses indicated the generalized validity of the MoCA scores. In conclusion, the results of this study show the overall psychometric adequacy of the MoCA and verify the discriminant and generalized validity of the obtained results.