To improve the utilization rate of clean energy and reduce energy losses, an energy dispatch strategy based on path optimality for multi-microgrid system is proposed. The single microgrid is composed of photovoltaic, wind power, storage battery, gas turbine and other equipment. Under the condition that multi-microgrid is connected to the grid, a two-layer optimal scheduling model of multi-microgrid system is established. The objective function of the first-layer scheduling is to minimize the maintenance cost, fuel cost, battery loss cost, pollutant treatment cost and power interaction cost, and the swarming spider optimization algorithm is adoted to solve the optimal output and total operation cost of each distributed generation unit in one cycle of the first layer. Based on the optimization results of the first layer scheduling, the second layer scheduling takes into account the energy scheduling losses between each microgrid and the large grid, and takes the minimum interaction cost between each grid as the objective function, adopts the ant colony algorithm to select the optimal path with the minimum loss, so as to realize the energy exchange among the multi-microgrids. Finally, a grid-connected multi-microgrid system based on IEEE 9-node is used for energy scheduling verification. The results show that the proposed method can effectively reduce the energy loss in the process of power mutual supply of each microgrid, and save the cost. The network loss is reduced from 1 379 kW to 905 kW, and the cost is reduced from 17 578 yuan to 13 443 yuan.