The sheep is an important domestic animal and model for many types of medically relevant research. An investigation of gene expression in ovine muscle would significantly advance our understanding of muscle growth. RNA-seq is a recently developed analytical approach for transcriptome profiling via high-throughput sequencing. Although RNA-seq has been recently applied to a wide variety of organisms, few RNA-seq studies have been conducted in livestock, particularly in sheep. In this study, two cDNA libraries were constructed from the biceps brachii of one Small-tailed Han sheep (SH) and one Dorper sheep (DP). The Illumina high-throughput sequencing technique and bioinformatics were used to determine transcript abundances and characteristics. For the SH and DP libraries, we obtained a total of 50,264,608 and 52,794,216 high quality reads, respectively. Approximately two-thirds of the reads could be mapped to the sheep genome. In addition, 40,481 and 38,851 potential coding single nucleotide polymorphisms (cSNPs) were observed, respectively, of which a total of 59,139 cSNP coordinates were different between the two samples. Up to 5,116 and 5,265 respective reference genes had undergone 13,827 and 15,684 alternative splicing events. A total of 6,989 reference genes were extended at the 5’, 3’ or both ends, and 123,678 novel transcript units were found. A total of 1,300 significantly differentially expressed genes were identified between the two libraries. These results suggest that there are many differences in the muscle transcriptomes between these two animals. This study addresses a preliminary analysis and offers a foundation for future genomic research in the sheep.