A recent genome-wide association study elucidated that 4q25 was implicated in ischemic stroke, but subsequent studies showed inconsistent results. In order to get coincident conclusion, we investigated two SNPs (rs2200733, rs10033464) on chromosome 4q25 in 1,388 stroke patients and 1,629 controls from Chinese Han population and then performed a meta-analysis. Although we failed to detect any association between 4q25 and stroke in our case-control study, meta-analysis revealed that rs2200733 showed association with overall stroke (OR 1.18, 95 % CI 1.08-1.27), but not for rs10033464. Subsequently subgroup analysis indicated that both rs2200733 and rs10033464 conferred increased risk for cardioembolic stroke (CE stroke) (for rs2200733, OR 1.38, 95 % CI 1.26-1.51; for rs10033464, OR 1.14, 95 % CI 1.02-1.26), while rs2200733 was marginal associated with non-CE stroke (OR 1.09, 95 % CI 1.02-1.16). our results demonstrated that two SNPs (rs2200733 and rs1003346) on chromosome 4q25 were limited to the stroke of cardioembolic etiology. To confirm this conclusion, well-designed studies with larger sample size involving case-control populations with homogeneous ancestry warrant to be conducted in the future.