DC power systems with multiple buses for redundancy are more reliable and provide reconfiguration options. A game-theoretic-based modeling approach for bus selection is proposed in this paper, which is based on local information of the player and does not require a centralized controller. The initial section provides the modeling and optimization of a single-input player. This modeling mainly follows on the players' local objectives and discrete choices of bus connections. Therefore, in this case, the controller obtains Nash equilibrium (NE) for the pure strategy game. Then, the approach is extended for the integration of global objectives with minimal communication. In addition, different objective priorities are integrated into the optimization routine. Next, the analysis is extended to multi-input player situations which are related to the class of mixed and continuous strategy games. This modeling is important, since the payoff matrix approach does not find an NE in all cases. Finally, the bus selection modeling is carried out by taking the system dynamics into account. This is important for the cases where the system has sudden load or source fluctuations. Experimental results are obtained, which validate the theory.