Objective The primary objective was to determine whether Danshen agents can improve functional outcome without causing undue harm in patients with acute ischaemic stroke. Secondary objectives were to assess the effect of Danshen agents on impairment and on the quality of life. Methods Searches were performed in the Cochrane Stroke Group Specialized Trial Register, Trials Register of the Cochrane Complementary Medicine Field, Chinese Stroke Trials Register and data from the pharmaceutical company. In addition, we searched the electronic bibliographic databases: Cochrane Controlled Trials Register (CENTRAL/CCTR) Issue 1, 2002, MEDLINE (1996 to 2002), EMBASE (1980 to 2002), China Biological Medicine Database (1978 to 2002). We handsearched ten Chinese journals potentially related to our question. Two reviewers selected studies, assessed quality of studies, extracted data independently. The primary outcomes of death or dependency at the end of long term follow-up(at least three months) and adverse events were assessed. Secondary outcome measures included: measures of neurological deficit at the end of treatment, death from all causes within the first two weeks of treatment and during the whole follow-up period and quality of life. Results Eight potentially eligible trials were identified, of which three trials (304 patients) were included. Two trials were excluded and three trials were waiting for assessment. Number of death and dependency at the end of long term follow-up (at least three months) were not reported in the three included trials. Only one trial reported the adverse events. Three trials measured neurological deficit at the end of treatment. Danshen agents were associated with a significant improvement in neurological deficit (RR 1.07, 95%CI 1.01 to 1.14). There was no death and during the whole treatment period and there was no assessment on quality of life. Conlusions There were too few patients and outcome events to draw reliable conclusions from the present data. The methodological quality of all included studies was poor. Further high quality randomised controlled trials should be performed.