Ischaemic stroke was confirmed by computerised tomography (CT) or magnetic resonance imaging (MRI) in 18 trials. Meta‐analysis of two trials with 164 participants showed that treatment with puerarin did not reduce death or dependency at final follow‐up (RR 0.79, 95% CI 0.45 to 1.36). One trial with 83 participants reported that the mean value of the Barthel Index in the puerarin group was below that in the control group. Meta‐analysis of 16 trials with 1305 participants showed that puerarin reduced the proportion of participants without improvement of neurological deficit at the end of follow‐up (RR 0.42, 95% CI 0.33 to 0.55). None of the included trials reported serious adverse effects.The quality of evidence was low due to incomplete reporting of the methods and short‐term follow‐up.