The future distribution network may encompass a large number of microgrids, in which case local networked microgrids can be formed and all the microgrids in the network be connected to the main grid through a distribution sub-station. To improve the efficiency of the entire network rather than focusing on improving the efficiency and reliability of each microgrid, this paper proposes a coalitional-game-theory-based local power exchange algorithm to identify incentives for coalitional operation and help microgrids in the network trade power locally with neighboring microgrids, so as to meet their own power requirements while achieving higher expected individual utility. Compared with the traditional operation situation, simulation results show that the proposed coalitional- game-theory-based local power exchange algorithm can help increase individual microgrid utility in the network. When there are 30 microgrids in the network, for example, each microgrid is expected to have a 16% increment of individual utility on average.