Although zebrafish was introduced as a laboratory model organism several decades ago and now serves as a primary model for developmental biology, there is only limited data on its genetic variation. An establishment of a dense polymorphism map becomes a requirement for effective linkage analysis and cloning approaches in zebrafish. By comparing ESTs to whole-genome shotgun data, we predicted >50,000 high-quality candidate SNPs covering the zebrafish genome with average resolution of 41 kbp. We experimentally validated approximately 65% of a randomly sampled subset by genotyping 16 samples from seven commonly used zebrafish strains. The analysis reveals very high nucleotide diversity between zebrafish isolates. Even with the limited number of samples that we genotyped, zebrafish isolates revealed considerable interstrain variation, ranging from 7% (inbred) to 37% (wild-derived) of polymorphic sites being heterozygous. The increased proportion of polymorphic over monomorphic sites results in five times more frequent observation of a three allelic variant compared with human or mouse. Phylogenetic analysis shows that comparisons between even the least divergent strains used in our analysis may provide one informative marker approximately every 500 nucleotides. Furthermore, the number of haplotypes per locus is relatively large, reflecting independent establishment of the different lines from wild isolates. Finally, our results suggest the presence of prominent C-to-U and A-to-I RNA editing events in zebrafish. Overall, the levels and organization of genetic variation between and within commonly used zebrafish strains are markedly different from other laboratory model organisms, which may affect experimental design and interpretation.