Expressed sequence taqs (ESTs) have accepted to be a valuable tool for discovering single nucleotide polymorphism (SNP) in many species. For detection putative SNPs in soybean genome, approximately 90,000 EST sequences from genotype Williams 82 were downloaded from NCBI. The paralog sequences of these ESTs were distinguished by TGI clustering tools (TGICL) performing megablast and EST cluster analysis, and the EST clusters were used as reference sequences for detection putative SNPs by in silico. The EST clusters were aligned with EST sequence from other cultivars of soybean by Polybase (computer software). The results revealed that putative SNPs were distributed in 5,677 clusters with frequency of 1 SNP per 333 nucleotide sites. For SNP validation, 43 primer pairs were designed from EST clusters containing putative SNPs for sequencing genomic DNA of Williams 82, Harosoy, Peking, Pureunkong, Jinpumkong 2, Hwangkeumkong and IT182932. From results of sequencing PCR, we totally found 99 SNPs from 33 primer pairs. Twenty-three and 47 out of 99 SNPs showed polymorphisms between Pureunkong and Jinpumkong 2, and Hwangkeumkong and IT182932, respectively. The SNPs discovered from this study can be used for genetic mapping in the four genotypes of soybean.