Unbalanced feeder systems cause deteriorating power quality and increase investment and operating costs. Feeder reconfiguration and phase swapping are two popular methods to balance the systems. For an unbalanced feeder system, feeder reconfiguration is difficult to meet the phase balancing criterion due to the limited number of sectionalizing switches available. Phase swapping is another alternative and direct approach for phase balancing. Phase swapping has not received its deserved attention due to the complexity of feeder systems, the dimension of problems, and totally overlooking the impacts of phase imbalance. Phase swapping can economically and effectively balance the feeder systems to improve power quality and reduce power system total cost. Deregulation arises competition on power quality, service reliability and electricity price. Therefore phase swapping can enhance a utilities' competitive capability. This paper proposes a mixed-integer programming formulation for phase swapping optimization. Single-phase loads are treated differently to three-phase loads. Nodal phase swapping and lateral phase swapping are also introduced. An example is used to illustrate the proposed method.