The　C4.5　Algorithm　can　result　in　a　thriving　decision　tree　and　will　overfit　the　training　data　while　training　the　model.　In　order　to　overcome　those　disadvantages,　this　paper　proposed　a　post-pruning　decision　tree　algorithm　based　on　Bayesian　theory,　in　which　each　branch　of　the　decision　tree　generated　by　the　C4.5　algorithm　is　validated　by　Bayesian　theorem,　and　then　those　branches　that　do　not　meet　the　conditions　will　be　removed　from　the　decision　tree,　at　last　a　simple　decision　tree　will　be　generated.　The　proposed　algorithm　can　be　verified　by　the　data　provided　by　the　Beijing　key　disciplines　platform　and　the　Beijing　Master　and　Dr.　Platform.　The　result　shows　that　the　algorithm　can　the　most　unreliable　and　uneven　branches.　And　compared　with　the　C4.5　algorithm,　the　proposed　algorithm　has　a　higher　prediction　accuracy　and　a　broader　coverage.　©　2013　IEEE.