An analysis of 101 stroke patients who were enrolled in 10 hospitals of Cities of Pusan, Taegu, and Andong from November 1, 1996 to April 31. 3997 was conducted using the modified Barthel Index(BI) and the adapted PULSES profile index(PS) to evaluate their function. Patients were examined at the following intervals: Initial assessment, one month after initial. at discharge, and one month after discharge. The mean BI score of patients initial assessment was 27.18, and that of PS was 17.54. There were statistically significant between initial score and one month after initial (21.39: p<0.001), at discharge(37.47: p<0.001), one month after discharge(46.49: p<0.001). PS scores were also improved .significantly(-2.62, -4.52. and -6.26(p<0.001). And the score between dischange and one month after discharge was significant (9.01: p<0.001) and in PS score(-1.73: p<0.001). Age and BE score were significantly associated with the improved in BI score between initial and discharge(T3-T1)(p<0.05). Below age forty the Bower initial BI score showed significantly higher improvement(T3-T1) after physical therapy(p<0.05). Initial Bl score, patients' attitude for physical therapy after discharge, age, and surgical operation were significantly associated with the improvement of BI score between initial and one month after discharge(T4-T1)(p<0.05). The lower initial BI score, patients' positive attitude for physical therapy after discharge, below age forty. and no surgical operation showed significantly higher improvement(p<0.05). BI score at discharge, side of hemiparesis and religion were significantly associated with the improvement of BI score between at discharge and one month after discharge(T4-T3) (p<0.05). The lower BI score at discharge. left aide of hemiparesis, with religion showed significantly higher improvement (p<0.05). Age, initial PS score were significantly associated with the improved in PS score between initial and discharge(T3-T1)(p<0.05). The higher initial PS score and below age forty showed significantly higher improvement(T3-T1)(p<0.05). Initial PS score, patient' attitude for physical therapy after discharge, age, educational level, physical therapy hour after discharge, and surgical operation were significantly associated with the improvement of PS score between initial and one month after discharge(T4-T1)(p<0.05). The higher initial PS score, patients' positive attitude for physical therapy after discharge, below age forty, higher education, the shorter physical therapy hour, and no surgical operation showed significantly higher improvement(T4-T1)(p<0.05). PS score at discharge, educational level, patient' attitude far physical therapy after discharge, physical therapy hour after discharge, and gender were significantly associated with the improvement of PS score between discharge and une month after discharge(T4-T3) (p<0.05), The higher PS non at discharge, higher education, patients' positive attitude for Physical therapy after discharge, the shorter physical therapy hour, and male showed significantly higher improvement (T4-T3)(p<0.05). In conclusion, initial BI score and age were significantly associated with BI score improvement and initial PS score, age, and educational level were also significantly associated with PS score improvement in stoke patients.