Aims The Cognitive Performance Scale (CPS), a minimum data set instrument of the interRAI, was initially designed to evaluate cognition in residential care and has demonstrated strong diagnostic accuracy. In this study, we evaluated the diagnostic accuracy and validity of the CPS in the post‐acute care setting among post‐stroke patients hospitalized in rehabilitation wards. Design Mixed methods. Methods The observational study was conducted in rehabilitation wards. Diagnostic accuracy was used to explore the level of agreement between CPS and Montreal Cognitive Assessment (MoCA) in 321 inpatients (62.12 years; 68.2% male). Results The diagnostic accuracy of the CPS was poor when MoCA was less than 24 as a gold standard, with an area under the curve of 0.69 (standard error 0.03, 95% confidence interval = 0.62–0.75). The CPS had a poor to moderate correlation with MoCA (rs = −.35).