The early identification of performance in the five-repetition sit-to-stand test (5-STS) at discharge in stroke patients could be of interest because it can determine independence for community-based activities. This study aimed to determine whether the initial measurement of the 5-STS test can be a determinant of the performance level prediction and amount of change in the 5-STS test at discharge in stroke patients. A prospective cohort study was conducted with a sample of 56 patients aged ≤60 d post-stroke. The 5-STS test results, as well as changes in patient condition, were measured at admission (T0) to an outpatient rehabilitation program, after the first month (T1), and at discharge (T2). The mean age was 62.7 (SD = 13.0), 58.9% of the subjects were male, and 75% had suffered an ischemic stroke. A multivariate linear regression model using the 5-STS test at T0 explained 57.7% of the variance in the performance at discharge. Using the 5-STS at T1 increased the variance to 75.5% (p < 0.001). Only the time from stroke onset at T0 significantly contributed to the two models. The measurement of the 5-STS at T0 and the amount of change in its performance at T2 explained 60.2% (p < 0.001) of the variance, while reassessment at T1 explained only 19.3% (p < 0.001). The level of patient performance on the 5-STS test at discharge, as well as any potential change, can be predicted by the admission measure of 5-STS in stroke patients.