分形块编码法能提供高压缩比、高质量的图象，其美中不足之处在于搜索最佳匹配块时耗时过长。本文提出的快速算法，采用二叉树和链表结构进行搜索，大大提高了匹配速度，从而缩短了编码时间。本算法分两次使用分形块编码，第二次是对第一次编码还原图和原图之间的误差进行编码，减少了图象失真。仿真结果表明，本算法较传统分形块编码法速度提高了几十倍。在高压缩比下，图象质量高于 Ｊ Ｐ Ｅ Ｇ算法。 Fractal block coding provides a high compression ratio, high quality image, and its downside is that it takes too long to search for the best matching block. The fast algorithm proposed in this paper, using binary tree and linked list structure search, greatly improves the matching speed, thus shortening the encoding time. The algorithm uses fractal block coding twice, and the second one is to encode the error between the original one and the original one to reduce the image distortion. The simulation results show that the proposed algorithm is more efficient than the traditional fractal block coding method by a factor of ten. At high compression ratio, the image quality is higher than J P E G algorithm.