To compare Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination (MMSE) global and subscores in classifying cognitive impairment in persons with mild stroke and to explore the relationship between admission and discharge functional status and improvement. Retrospective analysis of data. Acute rehabilitation unit of a large urban university-affiliated hospital. Inpatients with stroke (N=72; mean age, 70y; median time poststroke, 8.5d) and mild neurologic (median National Institutes of Health Stroke Scale score, 4) and cognitive deficits (median MMSE score, 25). Not applicable. Admission cognitive status was assessed by using the MMSE and MoCA. The motor subscale of the FIM instrument (mFIM) and motor relative functional efficiency was used to assess discharge functional status and improvement. The MoCA classified more persons as cognitively impaired than the MMSE (89% vs 63%, respectively; using a cutoff score of 27 on the MMSE and 26 on the MoCA). The MoCA also showed less of a ceiling effect than the MMSE, higher internal reliability (Cronbach α=.78 compared with α=.60), and marginally stronger associations with discharge functional status (r=.40; P<.001) than the MMSE (r=0.30; P<.05). The MoCA visuoexecutive subscore was the strongest predictor of functional status (P=.01) and improvement (P=.02) in global and subscores for both tests. The MoCA may be an important cognitive screening tool for persons with stroke and mild cognitive dysfunction on an acute rehabilitation unit. Lower visuoexecutive subscores may assist in identifying persons at risk for decreased functional gains in self-care and mobility (mFIM) during inpatient rehabilitation. The findings justify further validation studies of the MoCA in persons with subacute stroke. Copyright © 2011 American Congress of Rehabilitation Medicine. Published by Elsevier Inc. All rights reserved.