To　improve　the　accuracy　of　the　forecasting　of　the　college　building　energy　consumption,　this　pape　puts　forward　an　estimating　method　of　the　building　energy　consumption　according　to　the　grey　theory　and　radical　basis　function　neural　network　(RBFNN).　The　proposed　model　combines　the　advantages　of　low　data　demand　of　grey　theory　with　the　self-learning　and　self-organization　of　RBFNN.　Case　study　indicates　that　compared　with　those　of　the　traditional　grey　theory　and　RBFNN　models,　the　average　relative　deviation　between　predicted　and　the　real　value　can　decrease　5.　4%　based　on　the　proposed　model.