Abstract: It is a hot research area to substitute frozen trash fish (FF) with artificial compound diet (CD) in feeding the carnivorous economic fish. Largemouth bass (Micropterus salmoides) is one of the important aquaculture species with an annual output of about 340000 tons in fresh water in China. Effect of artificial compound diet at the transcriptome level of largemouth bass is rare reported. The current study utilized RNA-Seq (RNA sequencing) technology on the Illumina sequencing platform (Illumina HiSeq^TM 2000) to screen polymorphic SNP (single nucleotide polymorphism) markers of largemouth bass in two diets and analyzed the association of SNPs with growth trait. As a result, a total of 174M reads were assembled to generate 95024 unigenes. We used BWA (http://sourceforge.net/projects/bio-bwa/) and SamTools (http://source forge.net/projects/samtools/) to detect SNPs. Several filtering steps were performed to reduce the false positive of SNPs, 8681 SNPs met the selection criteria. 50 SNPs sequences in different genes were validated by PCR amplification and genotyped by the SNaPshot assay. 39 out of 50 SNPs were confirmed. Of these, 4 SNPs were monomorphic for all individuals and 35 SNPs were polymorphic. Validation of selected SNPs was 70%. 327 largemouth bass with marketable sizes fed by CD were used to evaluate the correlation between these validated SNPs and growth traits including body mass, total length, head length, body width, withers height and caudal peduncle length. Genotyping data were computed and analyzed by SPSS 19.0 statistical software package. Two SNPs were associated with body mass, total length and withers height, which could be used as important candidate molecular markers for breeding selection of largemouth bass. This study shows that RNA-Seq is an efficient approach to uncover gene-asso-ciated SNPs, which could facilitate genetic and functional genomics research in the artificial domestication of formulated feeding of largemouth bass.