Abstract: To improve the flexibility of microgrid operation, the interruptible load was regarded as a schedulable resource that directly participated in the operation of microgrid, and according to the length of time of peak load and valley load the transmission power of the tie line was set in different interval. Considering the depreciation, operation and maintenance cost of microgrid and its environmental benefit cost as well as its tie-line deviation penalty cost, a multi-objective optimization model of microgrid operation was constructed, and by use of relative objective adjacent scale method the proposed multi-objective optimization model was transformed into single-objective optimization model, and the combination weighting was performed by the analytic hierarchy process (AHP) and entropy weight method. The Gurobi solver was utilized to solve the 24h output and the state of charge of battery of various distributed power sources in the microgrid to obtain the day-ahead dispatching model of microgrid. By means of a simulation example based on a microgrid in a certain region in the United States, it was verified that the proposed model could give consideration to the economy and environmental protection property, besides, it could mitigate the power fluctuation in the main network, meanwhile it was shown that such an operation mode, which took the interruptive load and tie-line power control into account, can effectively reduce the depreciation cost of operation and maintenance, environmental pollution and the burden of main network.