OBJECTIVE：To investigate the correlat ion betwe en N-methyl-D-aspartate（NMDA）receptor subunit gene GRIN2B polymorphism and Han population with epilepsy in south Fujian. METHODS ：In retrospective study ，167 healthy people who had physical examination in the Affiliated Dongnan Hospital of Xiamen University were selected from Jan. 2017 to May 2018 as control group；163 epileptic patients who were monitored the blood concentration of sodium valproate were selected as epilepsy group. The clinical data and peripheral blood of 2 groups were collected. 12 loci of GRIN2B genotype（rs11055514，rs11055515，rs12814951， rs74816802，rs2160517，rs2193149，rs966664，rs1805476，rs1806201，rs1805522，rs3764030，rs1019385） in subjects were genotyped. Haploview 4.2 software was used to perform linkage disequilibrium （LD）analysis，and Pearson correlation was used to analyze haplotype. Distribution differences of wild homozygote （AA），mutant heterozygote （Aa）and mutant homozygote （aa） genotypes at 12 loci of GRIN2B gene between 2 groups were analyzed statistically by using GENO ，TREND，DOM and REC. Logistic regression model was used to analyze the correlation of epilepsy induction among 12 loci of GRIN2B gene. RESULTS ： Totally 12 loci of GRIN2B gene were all in line with Hardy-Weinberg equilibrium in 2 groups（P＞0.05）. There was an obvious LD phenomenon between the block 1 composed of rs 11055514，rs11055515，rs12814951，rs74816802，rs2160517，rs2193149 and rs966664 and the block 2 composed of rs 3764030 and rs 1019385（D’＞0.9，r2＞1/3）. There was a correlation between CGGACAG monoploid in block 1 and the occurrence of epilepsy （P＜0.05）. There was statistical significance in the distribution difference of rs74816802 and rs 2193149 between 2 groups（P＜0.05）. The mutation of rs 2193149 locus may cause epilepsy （addition and effect of alleles ：OR＝1.529，L95＝1.017，P＝0.041）. CONCLUSIONS ：The mutation of GRIN2B gene rs 2193149 locus may be one of the risk factors of epilepsy in Han population from south Fujian.