Data de-duplication eliminates multiple copies of the same file and repeated chunks or segments of data within those files, thus reduces disk capacity and bandwidth required for backups. Recently, driven by the needs of large-capacity/high performance data backup, data de-duplication is becoming a new research focus of the storage industry. In this paper, we present a fingerprint-based backup method named FBBM which performs data de-duplication in the backup. FBBM breaks files into variable sized chunks using anchor-based chunking scheme for the purpose of duplication detection. Chunks are stored on a write-once RAID (Redundant Array of Inexpensive Disks) and indexed and addressed by the hash of their content, this leads to intrinsically single instance storage. FBBM outperforms traditional backup methods in terms of storage and bandwidth saving. This makes it a promising tool for modern enterprises to backup or archive their ever-growing valuable data with low storage and bandwidth cost.