A general formulation and the development of a dynamic programming algorithm to solve a fuel-constrained unit commitment problem is presented. The system under consideration has oil units with storage constraints, gas units with delivery constraints, and unconstrained coal units. An optimal approach to transfer the fuel delivery constraints into unit capacity limits using a closed-form dual dispatch is introduced. As a result, the gas units can be viewed as fuel-unconstrained, thus greatly reducing complexity. It is shown that the optimization problem, which requires that storage levels be parameterized, leads to two different dispatch rules. The oil units are dispatched to maximize the power they generate for a binding fuel amount or cost, whereas the gas and coal units are economically dispatched to provide for the remaining power. In addition to the standard constraints, the system is required to meet a minimum system spinning reserve. Test results are provided to illustrate the merits of the proposed method