ABSTRACT: Objective: To investigate the efficacy and safety of a novel lower-limb exoskeletal robot, BEAR-H1 (Shenzhen Milebot Robot Technology), in the locomotor function of subacute stroke patients. Methods: The present study was approved by the ethical committee of the First Affiliated Hospital of Nanjing Medical University (No. 2019-MD-43), and registration was recorded on the Chinese Clinical Trial Registry with a unique identifier: ChiCTR2100044475. A total of 130 patients within 6 months of stroke were randomly divided into two groups: the robot group and the control group. The control group received routine training for walking, while in the robot group, BEAR-H1 lower-limb exoskeletal robot was used for locomotor training. Both groups received two sessions daily, 5 days a week for 4 weeks consecutively. Each session lasted 30 min. Before treatment, after treatment for 2 weeks, and 4 weeks, the patients were assessed based on the 6-minute walking test (6MWT), functional ambulation scale (FAC), Fugl-Meyer assessment lower-limb subscale (FMA-LE), and Vicon gait analysis. Results: After a 4-week intervention, the results of 6MWT, FMA-LE, FAC, cadence, and gait cycle in the two groups significantly improved (P < 0.05), but there was no significant difference between the two groups (P > 0.05). The ratio of stance phase to that of swing phase, swing phase symmetry ratio (SPSR), and step length symmetry ratio (SLSR) was not significantly improved after 4 weeks of training in both the groups. Further analyses revealed that the robot group exhibited potential benefits, as the point estimates of 6MWT and Δ6MWT (post-pre) at 4 weeks were higher than those in the control group. Additionally, within-group comparison showed that patients in the robot group had a significant improvement in 6MWT earlier than their counterparts in the control group. Conclusions: The rehabilitation robot in this study could improve the locomotor function of stroke patients; however, its effect was no better than conventional locomotor training.