Objective To evaluate the efficacy of Yangxue Qingnao Grain for migraine. Methods We searched The Cochrane Library (Issue 3, 2007), EMBASE (1974 to June 2007), PubMed (1966 to June 2007), VIP (1989 to June 2007), CNKI (1979 to June 2007) and CBM (1978 to June 2007) to identify randomized controlled trials or quasirandomized controlled trials comparing Yangxue Qingnao Grain versus sibeline for migraine. The quality of included studies was critically assessed and data analyses were performed with The Cochrane Collaboration’s RevMan 4.2.7. Results Eight studies were included, involving 5 randomized trials and 3 quasi-randomized trials. Meta-analyses showed that the total response rate (RR=1.07, 95%CI 1.00 to 1.15, P=0.06), the duration of headache (WMD 1.33, 95%CI – 0.87 to 3.52, P=0.24), and the frequency of headache episodes (WMD 0.93, 95%CI –1.00 to 2.86, P=0.35) were similar between Yangxue Qingnao Grain and sibeline. One study reported that the changes of ACA (WMD 3.70, 95%CI –3.46 to 10.86), MCA (WMD 0.60, 95%CI –10.37 to 11.57), ICA (WMD 3.40, 95%CI –4.35 to 11.15) and DCA (WMD – 2.30, 95%CI –9.52 to 4.92) revealed by transcranial doppler sonography were also comparable between the two groups. Conclusion Yangxue Qingnao Grain is effective in the treatment of migraine, but its superiority over sibeline is not demonstrated. Due to the poor quality and small sample of the included trials, more large-scale multi-center randomized trials are needed.