Background: Randomized controlled trials on early supported discharge (ESD) service in stroke have demonstrated favorable results. However, there are difficulties in transferring research results regarding ESD to the clinic. The aim of this study is to describe the method, content, implementation and outcome of ESD in its natural habit for stroke patients. Methods: A prospective observational implementation study of 153 consecutive stroke patients with mild to moderate severity with ESD and rehabilitation by a specially trained interdisciplinary team in the patient’s home, directly after discharge from the stroke unit. The interdisciplinary team in the ESD team is similar to the work at the stroke unit. Number of patients/year, clinical and functional health status, patient satisfaction, accidental falls/other injuries and resources were evaluated. Results: The number of patients/year in the ESD service has gradually increased from 2005 to 2009.The stroke patients subjected to Umea ESD service in January 2008 until May 2009 had a mean of 8.6 days of in-hospital care. The ESD service included 11 visits and 18 h per patient during 23 days (mean values). Compared with time of enrollment, patients exhibited reduced functional dependency (ADL- stairs 3 (1 - 5) vs. 1 (0–3), median, Q1–Q3, p <0.001, two-sided Wilcoxon Signed Ranks test) and increased mobility (Rivermead Mobility Index, (RMI) 11 (9–13) vs. 13 (12–15), p <0.001) at the time of discharge from the ESD service. Patient satisfaction regarding ESD was high. The long-term risk of accidental falls and other injuries appeared not to be increased. Conclusions: It is possible to locally develop and implement ESD care for stroke patients based on evidencebased principles. Our locally adapted ESD care, a stroke unit in the patients’ home, appears to be an appropriate alternative to conventional rehabilitation for patients with mild to moderate stroke.