Uneven energy consumption for wireless sensor network node will easily result in the problem of network partition, for which, a routing strategy based on game theory was proposed. Firstly, a game model between the sensor nodes with the residual energy as the utility was established, and then by solving the model, the optimal routing of the nodes remaining energy was ultimately determined. Simulation results show that the routing selection strategy can effectively maximize the nodes residual energy, and when the model is in Nash equilibrium, the energy consumption of each path is in equilibrium. So the energy resources of the sensor nodes can be efficiently used, avoiding the emptiness of network energy and prolonging the network lifetime.