In real-world applications, wireless sensor networks often consist of a large number of sensor nodes with constraint battery resources. How to reduce the power consumption of sensor nodes and maximize the network life, becomes the most important goal of topology control schemes in wireless sensor networks. During the operation of networks, sensor nodes may spend different levels of energy, and result in the uneven distribution of residual energy of sensor nodes. In order to extend the network life, it is essential to adjust the network burden of sensor nodes dynamically, so as to achieve energy balance among nodes under the consideration of different energy levels at nodes. In this paper, we introduce the game theory and the concept of game potential. By synthetically considering the factors of the residual energy and transmission power of nodes, a potential game based mathematical model of topology control is constructed. We prove the existence of Nash equilibrium. Through designing a payoff function, which takes into account both network connectivity and energy balance of nodes, the connectivity of sensor networks can be maintained while the power of sensor nodes is reduced. By increasing the average value of residual energy of neighbors, it enables to select nodes with more energy that reserves in neighborhood as neighbors, to improve the energy balance among nodes. Based on that, a distributed energy-balanced topology control algorithm (DEBA) is proposed. Theoretical analysis proves that the algorithm can maintain network connectivity. Compared with other existing game theory based algorithms DIA and MLPT, the topologies formed by the proposed algorithm have fewer bottleneck nodes which feature heavy traffic load and low residual energy, and smaller variance of node residual energy, thus achieving a longer life.