We present a new algorithm for compression of static color images. It allows to get higher compression ratios than the universal methods. Most of these traditional Lossless methods use techniques based on the elimination or reduction of the existent redundancy in the data (pixels), using mainly methods based on statistical models (e.g. Huffman Coding, Arithmetic Coding), dictionary models, pattern substitution (LZW…) or predictive coding of the adjacent pixel or near symbols (FELICS, JPEG…). This way, they lead to ratios which oscillate from 1:2 to 1:4 being considered these last ones as well acceptable results.Alternatively, we propose a new universal method based mainly on three sequential processes: first, segmentation of the image in fixed blocks, second, application of a compression algorithm based on the structure of data in form of binary tree and last, coding in order traversal of the binary tree. This method guarantees as minimum a ratio of compression of 1:3.Finally, we have applied the proposed method to a group of images of different sources and nature (photographic, satellite, medical etc.) and we have compared the experimental results with those given by the universal methods, among which is included JPEG Lossless proposed as an standard.