A method of compressing a 2×2 block based binary image is provided. The method includes: determining whether pixels included in a block are included in existing patterns; generating the number N of the pixels included in the existing patterns (N is a natural number); when N is more than a predetermined value, compressing the block with a binary bit stream comprising information about patterns of the pixels included in the existing patterns and color information about the pixels having a new pattern. The method of compressing a 2×2 block based binary image reduces information damage which may occur during compression and restoration of the binary pattern so that the difference in picture quality between an original image and the restored image can be visibly improved.