Nowadays, with the rapid development of the next generation sequencing technologies, metagenomics have become a new hotspot,However research in metagenomics faces the issue of binning --- identification and taxonomic characterization of the NGS short reads. To solve this problem, this paper first analyzes the next generation sequencing technology characteristics, statistical characteristics of metagenomic sequence, then proposes a new clustering method for DNA sequence binning. Test results show that this method has a very good clustering accuracy. In the same time, we developed an software for metagenomic binning based on this algorithm MetaBinning.