End-effector (EE) and exoskeleton (Exo) robots have not been directly compared previously. The present study aimed to directly compare EE and Exo robots in chronic stroke patients with moderate-to-severe upper limb impairment. This single-blinded, randomised controlled trial included 38 patients with stroke who were admitted to the rehabilitation hospital. The patients were equally divided into EE and Exo groups. Baseline characteristics, including sex, age, stroke type, brain lesion side (left/right), stroke duration, Fugl–Meyer Assessment (FMA)–Upper Extremity score, and Wolf Motor Function Test (WMFT) score, were assessed. Additionally, impairment level (FMA, motor status score), activity (WMFT), and participation (stroke impact scale [SIS]) were evaluated. There were no significant differences in baseline characteristics between the groups. After the intervention, improvements were significantly better in the EE group with regard to activity and participation (WMFT–Functional ability rating scale, WMFT–Time, and SIS–Participation). There was no intervention-related adverse event. The EE robot intervention is better than the Exo robot intervention with regard to activity and participation among chronic stroke patients with moderate-to-severe upper limb impairment. Further research is needed to confirm this novel finding.