OBJECTIVE: To investigate the effects of robot-assisted gait therapy on locomotor recovery in hemiparetic patients after stroke. METHOD: Twenty five stroke patients were randomly divided into 2 groups. Robotic training group received 30 minutes of robot-assisted gait therapy, 3 times a week for 4 weeks, while control group received conventional gait training with same duration and frequency as robotic group. Outcome was measured using Motricity index, Fugl-Meyer assessment (FMA), function ambulation category, modified motor assessment scale, 10-meter gait speed, isometric torque, Ashworth scale, gait analysis, body tissue composition, and Beck's depression inventory (BDI). RESULTS: Robotic training group showed significant improvement in motor functions measured by Motricity Index, FMA, 10-meter gait speed, isometric torque of hip compared with the baseline and with those of control group. Ashworth scale of hip, BDI, and muscle mass showed significantly greater improvement in robotic training group than control group. In gait analysis, stride length of unaffected leg demonstrated significant improvement in robotic training group (p<0.05). CONCLUSION: The robot-assisted gait therapy with body weight-support is considered to facilitate locomotor recovery of the hemiparetic stroke patient.