This paper proposes an efficient algorithm for considering reactive power and voltage constraints in unit commitment, Because the problem is formulated as a large mixed-integer programming and has a two-level hierarchical structure, the Benders decomposition method is used. Using decomposition, the master problem solves unit commitment, and the subproblem examines reactive power and voltage constraints in each hour to test the feasibility of the unit commitment solution. In the case of infeasibility, a proper constraint (Benders cut) is introduced and placed in the master problem for unit commitment rescheduling. The iterative process between the master and subproblems provides a minimum cost solution for generation scheduling while satisfying reactive power and voltage constraints. The constrained economic dispatch is used in our unit commitment. The constraints in the subproblem are formulated as linear programming with a block-angular structure. So, the Dantzig-Wolfe decomposition is applied in the subproblem to solve large-scale power system problems. A modified IEEE-30 bus system is presented to demonstrate the efficiency of the method.