Abstract: Local texture was the basic element of recognizing the whole image and starting with analysis of local texture was a reasonable way in developing better decorrelation model. Based on this idea, a new lossless image data compression algorithm named lossless image compression based on feature of local texture (FOLT) was presented. The algorithm was composed from two parts, decorrelation and entropy coding. Decorrelation was based on adaptively selection best decorrelating model suitable for the processed local area from a given model group. And the CCSDS adaptive entropy coder was used for entropy coding. The compression results not only show that the performance of the proposed algorithm is better than that of JPEG-LS,but also show the proposed decorrelation idea is effective.