Zebrafish has been widely used for the study of human diseases, as ~70% of the protein-coding genes are conserved between the two species. Moreover, zebrafish embryos are transparent and thus can serve as an ideal model for genetic studies in animal development. Surprisingly, the functional annotation of zebrafish genome itself has been severely lagging when compared with other model systems such as mouse and drosophila. Here we took a similar approach adopted by the ENCODE and Roadmap Epigenomics projects, and performed RNA-Seq and ChIP-Seq for several histone modifications to generate a comprehensive map of transcriptomes and regulatory elements in a variety of zebrafish tissues, including brain, heart, liver, skeletal muscle, kidney and several embryonic tissues. We predicted over 100,000 cis-regulatory elements in the zebrafish genome, the most comprehensive functional annotation effort in zebrafish so far to our knowledge. We also identified tissue-specific and developmental stage-specific regulatory elements. By comparing the data generated by the ENCODE and Roadmap Epigenomics projects, we also defined a set of functionally conserved and species-specific regulatory sequences among zebrafish, mouse and human. In summary, we generated a great genomics/epigenomics resource for the functional annotation in the vertebrate genomes and further expanded the value of zebrafish as a model of human disease.