BACKGROUND AND PURPOSE: Treadmill training with partial body weight support is a new and promising therapy in gait rehabilitation of stroke patients. The study intended to investigate its efficiency compared with gait training within regular physiotherapy in nonambulatory patients with chronic hemiparesis. METHODS: An A-B-A single-case study design compared treadmill training plus partial body weight support (A) with physiotherapy based on the Bobath concept (B) in seven nonambulatory hemiparetic patients. The minimum poststroke interval was 3 months, and each treatment phase lasted 3 weeks. Variables were gait ability assessed by the Functional Ambulation Category, other motor functions tested by the Rivermead Motor Assessment, muscle strength assessed by the Motricity Index, muscle tone rated by the Modified Ashworth Spasticity Scale, and gait cycle parameters. RESULTS: Treadmill training was more effective with regard to restoration of gait ability (P < .05) and walking velocity (P < .05). Other motor functions improved steadily during the study. Muscle strength did not change, and muscle tone varied in an unsystematic way. The ratio of cadence to stride length did not alter significantly. CONCLUSIONS: Treadmill training offers the advantages of task-oriented training with numerous repetitions of a supervised gait pattern. It proved powerful in gait restoration of nonambulatory patients with chronic hemiparesis. Treadmill training could therefore become an adjunctive tool to regain walking ability in a shorter period of time.