In　this　paper,　a　approach　for　automatically　generating　fuzzy　rules　from　sample　patterns　is　presented.　Firstly,　with　Cauchy　mute　operator　and　Gaussian　mute　operator,　we　propose　a　new　evolutionary　programming(EP)　based　on　self-adaptive　EP.　Secondly,　a　self-adaptive　fuzzy　neural　network　is　built　based　on　the　new　evolutionary　programming.　In　this　method,　structure　identification　and　parameters　estimation　are　performed　automatically　and　simultaneously.　The　simulation　results　show　that　the　proposed　method　in　this　paper　can　produce　the　compact　and　high　performance　fuzzy　rule-base　in　comparison　with　other　algorithms.　©　2007　IEEE.