Objective: To investigate the effects of training using a trunk control robot (TCR) system combined with conventional therapy (CT) on balance and gait abilities in persons with chronic stroke. Design: Two-group pretest-posttest design. Methods: Thirty-five subjects with chronic stroke were randomly assigned to either the TCR group (n=17) or the trunk extension-training (TET) group (n=18). Both groups performed CT for 30 minutes, after which the TCR group performed TCR training and the TET group performed trunk extension training for 20 minutes. Both groups performed the therapeutic interventions 3 days per week for 6 weeks. Balance ability was evaluated using the Berg Balance Scale (BBS), and the Timed Up-and-Go (TUG) test. Gait ability was measured using the 10 m Walk Test (10MWT) and the NeuroCom Smart Balance Master. Results: TCR group showed significant improvements in static balance (weight bearing) and dynamic balance (weight shifting speed, weight shifting direction, BBS, and TUG), 10MWT, gait speed, and step width (p<0.05); step length was not significant. The TET group showed a significant partial improvement of dynamic balance (weight shifting speed, weight shifting direction, BBS, and 10MWT (p<0.05), but the improvements in static balance, TUG, gait speed, and step width and step length was not significant. Additionally, significant differences in static balance, dynamic balance (weight shifting speed, weight shifting direction, BBS, and TUG), 10MWT, gait speed, and step width were detected between groups (p<0.05). Conclusions: TCR training combined with CT is effective in improving static and dynamic balance, as well as gait abilities in persons with chronic stroke.