Background: The time lapse between the acute event and the beginning of rehabilitation seems to play a significant role in determining the effectiveness of rehabilitation together with the severity of neurological deficits and impairment of motor and cognitive functions. The present study aims to further explore the prognostic role of cognitive and motor functions, concerning the different times of the beginning of neurorehabilitation. Methods: A secondary examination was conducted by applying a cluster analysis on the data of 386 stroke patients in the subacute phase who were enrolled in the Cognitive and Recovery of Motor Functions (CogniReMo) study. Results: Barthel Index at the admission predicts clinical outcome: if BI was 0 it was on average 28.7±24.1 at discharge. For patients with Barthel Index <15 at discharge, the discriminant was unaltered executive functions to have an average output of 61.3 instead of 45.5. In the range of BI at admission between 16 and 45 the discriminant variable was to have a NIHSS≤5 to obtain a high outcome (BI=75.4 instead of BI=61.9). Subjects with a BI at admission >45 were the best responders to rehabilitation, with a mean BI at discharge of 85 if they have alteration in spatial attention, and 95.3 if they have no deficits in spatial attention. Also, for in-patients hospitalized in a period ranging from the 20th to the 37th day after stroke, spatial attention was a discriminant variable to have a poor outcome (BI=34.3) vs. a good one (BI=76.7). Conclusions: the algorithm identified a hierarchical decision tree that might assume a significant role for clinicians in defining an appropriate rehabilitation pathway, depending on the time of rehabilitation beginning and the severity of motor and cognitive deficits.