Single nucleotide polymorphism(SNP) markers coupled with the next-generation sequencing technology are powerful tools for studying population genetics, as well as for species delimitation. The popular ways of obtaining SNP data include RNA sequencing and restriction site associated DNA(RAD) sequencing. However, both methods have limitations. For example, RNA sequencing requires fresh tissue samples, whereas RAD sequencing targets anonymous loci usually not transferable across species. In comparison with these methods, target gene capture could be used to enrich homologous fragments across divergent species and even from degraded DNA. We showed that by targeting a set of universal single-copy nuclear gene markers of vertebrates, we could retrieve thousands of SNPs from the Japanese grenadier anchovy(Coilia nasus) sampled from three locations:coastal region close to the estuary of the Yangtze River, the mainstream of the river and Lake Dongting, a lake connected to the middle reaches of the Yangtze River. We analyzed hundreds of representative SNPs from the data using STRUCTURE and Bayes factor delimitation(^*with genomic data; BFD^*), a new Bayesian species delimitation tool. We found that the fish from Lake Dongting are genetically different from the fish of other sample locations. We also observed marginal difference between fish collected from the coastal region and the mainstream of the Yangtze River(Bayes factor=11.3). Our study demonstrated that gene capture could be used to generate SNP data for species delimitation applying new analytic tools such as BFD^*.