The receiving of good coefficients of the image compression probably in the event that the compressor is specialized and it takes into account special features of the image structure. The submitted method of the statistical block compression of images takes into account the presence in the image many areas with almost identical brightness. The in itself given method does not represent the special value, as its unique positive quality is account of special features of the image structure. The method makes the analysis of the image on with the purpose of detection of areas with close brightness. But the given method actually reduces volume of an initial file, coding the information thus, at which it is convenient for compression by a method LZW. The method of statistical block compression of images consists in the analysis of blocks, which turn out by division of the entrance image. The size of received blocks is identical. Above received blocks are made affine transformations and those blocks, which satisfy to the element control criterion are considered identical. The basic task consists in search of the greatest amount of acceptable blocks. Certainly, the search is possible to make very much plenty of time, as if we have the image size 100 X 100, the amount of blocks of the different size can be equal 10,000. But to make it is unprofitable, therefore block has a kind of a square. The submitted method has two variants: one uses the block of the fixed size 8 X 8 (given size was determined experimentally), another uses blocks of the various size, which turn out by dynamic division of unacceptable blocks. Upon termination of the analysis of blocks at presence of acceptable blocks in a final file the following information enters the name: amount of blocks, blocks, coded information, which defines the place of the block and the identifier of affine transformation. Two variants of the submitted method practically are identical on efficiency, as use of dynamic division of the block improves the factor of compression on 15 - 20%, but increases the program time.