Abstract: Because of the situations that the uncertainty of system, the interconnection among multi integrated energy system (EIS) and interaction among electric, heating and cooling system are not given full consideration, optimal planning of EIS at user side based on bi-level programming is studied in this paper. A method generating multi-state of PV-load in different period is presented by adopting density clustering. Considering the interconnection of multi energy as well as multi-energy router coordination and the combined dispatch of multi energy flow, a bi-level programming model containing is then presented. The elitist strategy genetic algorithm is employed to solve the upper model and branch & bound algorithm is applied to optimizing lower model. The simulation results indicate: the proposed planning method of energy internet is effective.