Mei flower (Prunus mume) belongs to the Rosaceae family, has been domesticated for more than 7000 years and now is widely cultivated in East Asian countries as the famous tree provides both ornamental flowers and tasty fruits which can tolerate temperature below zero in early spring. We generated 50.4 Gb of sequencing data using the whole-genome shotgun (WGS) sequencing strategy with Illumina Genome Analyzer sequencing technology to sequence 3 short insert (180 to 800 bp) and 5 large insert (2 to 40 kb) size libraries. Using the stringent criteria in order to reduce the effection of sequencing error to the assembly, 28.4 Gb data were considered for de novo genome assembly. SOAPdenovo, a genome assembler algorithm based on de Bruijn graph, was used to assemble the 237.1Mb Prunus mume genome with a contig N50 31.8 Kb (longest 201.1 Kb) and scaffold N50 577.8 Kb (longest 2,871.0 Kb). In addition, to improve assembly quality, we carried out optical mapping technology and generated 83.7 Gb data and the final scaffold N50 increased to 1.2 Mb. We used a combination of de novo gene prediction programs and homology-based methods together with RNA-Seq data to predict gene models in the Prunus mume genome. After combined the different gene models, we identified 31,537 genes with average transcript length of 2,853 bp, coding sequence size of 1,146 bp and 4.61 exons per gene. Together with other sequenced species in Rosaceae, we should make a further step in describing the evolution of the Rosaceae.