2-D　maximum　entropy　algorithm　is　of　quite　good　effect　in　image　segmentation,　but　it　requires　long　time　on　the　complex　calculation.　Considering　CGA＇s　(Chaos　Genetic　Algorithm)　ability　to　retain　the　species　diversity　and　great　astringency,　a　new　2-D　maximum　entropy　method　based　on　CGA　was　put　forward.　It　has　been　proved　that　the　new　algorithm　is　of　better　capacity　to　search　for　the　best,　performs　more　steadily　and　results　in　better　segmentation　effect.