The various assessments performed by rehabilitation professionals not only indicate the patient’s current functional status but can also help determine the future status (prediction) or the ability to perform untested tasks (discrimination). In particular, the cutoff values are the simplest predictive and discriminative tool that can be widely used in clinical practice. The purpose of this scoping review was to summarize the current literature on cutoff values of motor and cognitive function for predicting or discriminating levels of activities of daily living after stroke. A literature search was conducted using the PubMed, CINAHL, and Scopus databases. The creation of the search criteria, primary screening of titles and abstract, and secondary screening by full-text review were performed by two rehabilitation professionals. A total of 54 articles were included. The summary of the cutoff values for prediction based on longitudinal studies revealed that an NIHSS score ≤8 and mRMI score ≥19 at acute hospitalization can predict good functional independence and walking independence, respectively, indicating reliable cutoff values. Cutoff values for predicting specific ADLs, such as toilet use or dressing, were not reported, which was a potential research gap identified in this review. Alternatively, the summary of the cutoff values for discrimination based on cross-sectional studies revealed that 288–367.5 m on the 6-min walk test and 25.5–27.6 points on the FMA-LL can discriminate community and noncommunity walkers. Considering the difference between prediction and discrimination, the reliable predicted cutoff values revealed in this review are useful for planning an intervention based on early prediction. Conversely, cutoff values for discrimination can estimate different performances with simpler test, or use as target values during rehabilitation.