This　paper　proposes　a　prediction　approach　on　energy　consumption　for　public　buildings　based　on　mind　evolutionary　algorithm　and　BP　neural　network.　The　actual　real-time　data　of　some　layer　in　a　public　building　can　be　obtained　online　by　our　implemented　building　monitoring　system,　then　several　key　factors　which　affect　building　energy　consumption　can　be　analyzed　and　determined　by　correlation　analysis　method.　By　using　the　mind　evolutionary　algorithm,　the　ideal　weight　values　and　threshold　values　of　BP　neural　network　are　calculated,　which　can　solve　its　problems　of　low　efficiency　and　slow　convergence.　Finally,　the　performance　and　effectiveness　of　the　proposed　forecasting　model　are　demonstrated　through　a　case　study　of　a　building　energy　consumption　monitoring　system　from　practical　engineering.