The game model of dynamic island partition for distribution network with DGs is proposed on the basis of the maxi-mum load recovery and the stability of distribution network. In addition, the game algorithm of dynamic island partition is also designed. Firstly, the distribution network of different time among islanding is regarded as different gamers and the strategies of the game are also defined. Secondly, the indexes consist of the amount of load recovery, transmission loss and numbers of switching operations are designed and the revenue function under non-cooperative and cooperative games is established. Then, the game algorithm of dynamic island partition and the application of Quantum Particle Swarm Optimization (QPSO) in the game iterative process are both proposed. Finally, the strategies of island partition among different islanding time are calculated by the simulation of two scenarios based on IEEE 33-bus distribution system. The conclusion that different islanding time may make a different effect on island partition strategy and the effectiveness of the proposed algorithm in solving the island partition are validated through the analysis of two scenarios. This work is supported by National Natural Science Foundation of China (No. 51147010 and No. 51207139).