Background: Rainbow trout is important fish species for aquaculture and biomedical research but has no genomic data. Until genome sequence becomes available, transcriptome sequencing is a rapid and efficient means for gene discovery and genetic marker development. Large-scale EST (258,973) Singer sequences are publicly available for rainbow trout. However, the nature of duplicated rainbow trout genome hinders assembly and annotation of the EST sequences. Additionally, previous efforts aimed at SNP discovery for rainbow trout using ESTs were unsuccessful, mainly, due to difficulties parsing allelic variation from the high frequency of duplicated genes. Results: High-throughput shotgun deep sequencing of the rainbow trout double-haploid transcriptome using DNA 454-pyrosequencing technology has been successfully applied yielding about 1.3 million reads with an average length of 344bp, a total of 447 million bases. De novo assembly of the sequences yielded 151,847 Tentative Consensus sequences (TCs) (Average length 662 nt) and 224,391 singletons. A combination assembly of both the 454-pyrosequencing ESTs and the pre-existing Singer sequences resulted in 161,818 TCs (Average length 758 nt) and 261,071 singletons. Gene Ontology analysis of the combination assembly showed similarity to the expected transcriptome of other fish species with known genome sequences, suggesting a genome-wide representation of the rainbow trout transcriptome sequence. Conclusion: The 454 library added great amount of new EST sequences and identified new genes. In addition, it improved assembly and annotation of the rainbow trout Sanger EST. The 454 library is a new tool for functional genome research in rainbow trout. It provides a reference sequence to identify gene duplications, allelic variations distinguish true/false SNPs as well as for digital gene expression and proteomic research in rainbow trout.