Objective: We aimed to compare the responsiveness, concurrent and predictive validity of the shortened Fugl-Meyer Assessment (S-FMA) and the streamlined Wolf Motor Function Test (S-WMFT) in persons with subacute stroke. Design: Test–retest design. Setting: Departments of physical medicine and rehabilitation at three hospitals. Participants: Participants with first-time stroke ( N = 51; 38 men, 13 women; mean age ± SD, 55.1 ± 11.7 years) based on scores of Mini-Mental State Examination and Brunnstrom stage. Interventions: Participants received one of three rehabilitation therapies for three weeks and were evaluated at baseline and end of treatment. Main outcome measures: Responsiveness was examined using the paired t-test and the standardized response mean (SRM). Criterion validity was investigated using the Pearson’s correlation coefficient ( r). Results: Changes from baseline to end of treatment assessed by both tests were significant ( P < 0.001). The value for responsiveness of the S-FMA was significantly higher than that of the S-WMFT (SRM difference, 0.48; 95% confidence interval, 0.23–0.63). There were stronger associations between the comparison scales and the S-FMA ( r = 0.57–0.68) than with the S-WMFT ( r = 0.39–0.58). Conclusions: The S-FMA had better concurrent and predictive validity than the S-WMFT and was more sensitive to changes caused by rehabilitation therapies. The S-FMA is recommended for expedited assessment of arm motor function outcome in stroke patients receiving rehabilitative therapy.