The genetic structure of species plays an important role in inferring population dynamics such as effective population size, geographical distribution, gene flow and population genetic differentiation. In order to reveal the genetic diversity level and genetic structure of Leiognathus equulus populations, a total of 92 individuals from 3 populations were collected and analyzed by control region fragments. The length of the control region fragment was 393 bp. A total of 32 haplotypes were detected for 92 individuals and 7 haplotypes were shared. The haplotype and nucleotide diversity of Xinzhu population was lower than that of two populations from Hainan. The whole haplotype and nucleotide diversity was 0.61±0.12 to 0.86±0.05 and 0.003 3±0.002 4 to 0.005 3±0.003 4, respectively. Two clades (Clade A and Clade B) were obtained in the neighbor-joining tree and median-networks. Clade A consisted of 22 haplotypes, which all were from Hainan populations; Clade B consisted of 10 haplotypes, which all are from Taiwan except Hap13. The haplotype and nucleotide diversity of Clade A was 0.81±0.04 and 0.003 7±0.002 5, and the haplotype and nucleotide diversity of Clade B was 0.69±0.10 and 0.003 3±0.002 4. Significant genetic differentiation was detected between Xinzhu population and the other two populations. The results of AMOVA showed that most genetic variation occurred among groups, which account for 75.37%. Tajima’s D and Fu’s F_s statistics for two clades were significantly negative for these two clades, which rejected the hypothesis of selective neutrality. The mismatch distribution of L. equulus appeared to be unimodal for two clades, and closely matched the expected distributions under the sudden-expansion model. The expanding time may be nearly 52 500-105 000 and 67 600-135 200 years ago in the late Pleistocene for two clades. The present phylogeographic pattern of L. equulus populations may be the combination of historical factors and current factors. The isolation of the Pleistocene glacial land bridge resulted in genetic differentiation of the populations and the isolated population had a secondary connection. Geographical distance hinders gene exchange and promotes the formation of the existing distribution pattern of L. equulus populations.