one of the common characters between traditional algorithms based on dct image coding is that regardless of the content of the image, they use a fixed size block to process, for example, jpeg, mpeg-1 / 2, h . 263 use constant 8 * 8 block mode . the main work of this article is to propose a sequence images compression algorithm that is based on adaptive block dividing and object to the disadvantages of the one based on fixed-size block dividing . the proposed algorithm is an image compression method, which can auto-adjust dct block size ( 8 * 8 or 4 * 4 ) according to the difference of corresponding blocks between two frames of the sequential image and doesn ’ t process any unnecessary data blocks in order to enhance coding efficiency 传统的基于dct的图像编码算法的一个共同特点就是:不管图像的具体内容,采用固定的块尺寸进行统一的处理,例如jpeg,mpeg-1/2,h.263均采用固定的88分块方式,本文的重点工作是针对这些采用固定的块尺寸进行统一处理的算法的一些缺点,提出了一种基于自适应分块的序列图像压缩算法,该算法是一种根据序列图像两帧间对应块的差别自动调整dct变换尺寸(88或44),对不必要的数据块不进行处理,以提高编码效率的序列图像压缩编码方法,具有较高的编码效率和较好的压缩效果,与传统的采用88固定分块方式的压缩编码算法相比,性能有较大幅度的提高,在文中,作者对两者的压缩编码性能进行了分析比较。