As　a　new　technology　of　renewable　energy,　ground-coupled　heat　pump　system　has　been　rising　in　our　country　in　recently　years.　However　because　of　the　nonlinear　and　high　degree　of　coupling,　it　is　not　clear　to　find　out　the　relationship　between　control　variables　and　total　energy　consumption　of　the　system.　This　paper　proposes　a　modeling　way　for　ground-coupled　heat　pump　system　that　uses　artificial　neural　network,　it　builds　the　relationship　between　control　variables　and　total　energy　consumption　of　the　system　directly.　Through　the　comparison　we　can　find　a　better　way　for　building　the　Ground-coupled　heat　pump　System.