Abstract: A partial genomic library of Pseudosciaena crocea was constructed in this paper. The PCR-based screening library method was used to isolate microsatellite loci containing the repeat motif of (CA)n and (GA)n. Total 55 positive clones were isolated. 32 positive clones were sequenced. 8 pairs of primers were designed based on the unique sequences flanking each motif with the software Primer3. Genomic DNA from 92 tested individuals was applied to evaluate the amplification efficiency of these primers and polymorphism of these markers. GENPOP (v.1.32) software was performanced to analyse the allele data. The number of the alleles per locus varied from 6 to 22. Mean observed and expected heterozygosities(Ho and He),were 0.404 7 and 0.775 6, respectively. These markers showed polymorphism information content between 0.365 1 and 0.928 1.Four P.crocea stocks from different geographic regions were served to evalucate the population genetic structure with these microsatellites DNA logi. The results showed that genetic differentiation were not significant between all stocks (Fst=0.084 8).FIS values within each stock were something high (average 0.436 7) and tests for all sampled populations contained unique alleles. The findings indicated that the populations of P. crocea still possessed variation between tested stocks.It may be caused by the large geographical scales of sample stocks. It means if conducting the suitable conservation and management strategies, the germplasm of large yellow croaker could renew in some measure. The observed allele frequency of most of microsatellites loci deviated from Hardy-weinberg equilibrium, i.e. population mating system may not follow the random principle. The gene flow between all four stocks were 2.699 8, data implied that there were significant gene exchange between stocks. Clustering analysis results supported that P. crocea from China's coast would divided into three different geographic populations.