High-sensitivity C-reactive protein (hsCRP) was reported as a strong, independent predictor of future myocardial infarction and stroke. It is of importance to illustrate the conformance of CRP genetic variation, increment of plasma hsCRP and cerebral events. A case-control study including 548 patients with acute ischemic stroke and 993 age-matched controls from community-based population was conducted and four tagging SNPs (tagSNPs) were genotyped. Multiple logistic regression was applied to evaluate the association of CRP gene and stroke hsCRP elevation with adjustment for covariates. The results indicated that rs3093059 and rs3091244 presented statistical associations with ischemic stroke. Odds ratios (ORs) (95 % confidence interval [CI]) of additive model, dominant model and minor allele at rs3093059 were 0.697 (0.528-0.921), 0.671 (0.487-0.923) and 0.811 (0.666-0.988), and ORs (95 % CI) of dominant model at rs3091244 was 0.728 (0.536-0.988), after adjusting for covariates. But there were no significant differences of genotype or allele frequencies of the four SNPs observed between hypertension (HT) and normal blood pressure (NBP) groups. Further analyses indicated the genetic variations of rs876537 and rs3093059 were positively associated with increased square root transformed hsCRP and hsCRP elevation (≥3 mg/l) in ischemic stroke patients, and rs876537 and rs3091244 were associated with hsCRP elevation in controls as well. Our finding suggests that the CRP genetic polymorphisms were associated with decreased risk of ischemic stroke and elevated plasma hsCRP and further replication study and functional research would be warranted.