BACKROUND AND PURPOSE: Our goal was to determine effects of partial body weight-supported(BWS)-treadmil training on subacute stroke patients. METHODS: Twenty patients were randomly assigned into two groups.Patients in groupl were participated in conventional treatment and partial BWS-treadmil training.Patients in group2 were participated only in conventional treatment.Functional Ambulation Scale(FAS),Rivermead Motor Evaluation gross(RMD1) and total gross(RMD2) function, Berg Balance Scale,Barthel Index(BI), walking-distance(6min), walking-time(10m), cadence rate, ratios of right-left step length,s(EMG), superficial muscle activity were evaluated in pretreatment and posttreatment periods, and 3-months. RESULTS: Patients in group1 showed statistically significant improvements in all parameters at posttreatment and 3-months(p<0.05).In group2, while there were significant improvements in FAS, RMD2, TA, gastrocnemius s(EMG) measurements at posttreatment and 3-months(p<0.05), no improvements were found in ratios of right-left step length at posttreatment period(p>0.05); for remaining parameters, improvements were found at 3-months(p>0.05). When groups were compared with regard to differences between pretreatment and posttreatment and pretreatment and 3-months, with being in favor of group1, except for 10m walking time, there were significant differences FAS at posttreatment period and 3-months(p<0.001); RMD1 at posttreatment period(p<0.001) and 3-months(p<0.05); RMD2 and Berg Balance Scale at posttreatment period and 3-months(p<0.01); 6min walking-distance cadence and TA s(EMG)at posttreatment period(p<0.01) and 3-months(p<0.05); gastrocnemius s(EMG) and ratios of right/left step length only at posttreatment period(p<0.05). CONCLUSION: Partial BWS-treadmil training combined with conventional treatment was found to be more effective on walking training compared to only conventional treatment.