The large-scale penetration of renewable energy has promoted great attention on the DC microgrid. For the supply of new energy, the DC microgrid can reduce loss and instability of AC/DC conversion. Therefore, in the DC microgrid environment, a multi-master-slave game model which regards the sellers as the leaders and the users as the followers is constructed. Then a game analysis for the seller’s price and user’s demand is carried out and finally multiple seller-user pairs can be obtained according to the distributed selection algorithm. Based on two routes with different transmission losses of channels, the optimal power packet scheduling of multiple matched pairs is realized by genetic algorithm. Simulation results demonstrate the feasibility and superiority of the two proposed algorithms. This work is supported by National Natural Science Foundation of China (No. 61501185) and Science and Technology Project of State Grid Hebei Electric Power Co., Ltd. (No. SGHEHS00FCJS1800274).