A　mass　of　engineering　data　is　induced　to　some　experiential　formulas　via　data　modeling　method,　which　substitutes　original　data　to　compute　and　stand　for　trend　to　data　in　certain　degree.　Computation　examples　from　numerical　results　of　complex　Reynolds　equations　and　the　measured　data　in　the　viscosity-pressure　experiment　show　that　data　modeling　method　based　on　evolutionary　algorithm　is　effective　for　automation　of　data　modeling　process　with　high　accuracy.　This　method　is　employed　to　analyze　the　hidden　rules　and　to　discover　simple　mathematical　models　from　a　great　deal　of　characteristic　data　of　slider　bearings.　It　is　practical　and　convenient　for　systematical　analysis　and　engineering　applications.