How　to　solve　diagnosis　problem　effectively　by　the　aid　of　Artificial　Intelligence　has　been　being　a　focus　of　Intelligent　System　researching　and　development.　However,　performance　of　the　native　Uncertainty　Reason　method　still　is　unsatisfactory　to　professionals　and　researchers.　This　paper　proposes　the　concept　of　Pivot-Association　Factor　and　its　Certainty　Index　Correction　algorithm,　and　on　this　base　designs　an　innovative　AI　method　for　solving　disease　diagnosing.　By　the　prototype　system　of　the　novelty　diagnosing　algorithm,　researcher　team　practises　a　series　of　test　experiments.　Results　and　analysis　demonstrate　that　the　method　exhibit　a　better　diagnosing　efficiency　than　man-expert　individual　and　the　native　Certainty　Factor　Diagnosing　Method.　On　the　whole,　it　manifests　an　accurate　rate　of　diagnosis　over　82%　and　possesses　the　capacity　to　expand　diagnosing　capability　of　man-expert.　©　2013　Binary　Information　Press.