To achieve the energy consumption reduction of buildings, the energy operation strategy of combined cooling, heating and power (CCHP) system, thermal storage and battery shared by buildings in the building clusters is addressed. The energy demands and energy prices are stochastic, and each building aims to minimize its cost. The stochastic programming approach and multi-objective theory are applied to formulate a two-stage multi-objective stochastic programming model. In order to obtain the Pareto optimal solutions, the LP-Relaxation and Benders decomposition algorithm are adopted to solve the energy schedule model with high solving efficiency. Finally, the mathematical model is solved by CPLEX software, and comparative studies are performed to evalute the effect of the different stochastic parameters on the optimal building cluster total cost and every building cost. The experimental results show that the hybrid algorithm is effective for solving the optimal model and the multi-objective stochastic programming model can significantly improve the accuracy of the optimal decisions.