This paper proposes a new algorithm—binaryglowworm swarm optimization (BGSO) to solve the unitcommitment (UC) problem. After a certain quantity ofinitial feasible solutions is obtained by using the prioritylist and the decommitment of redundant unit, BGSO isapplied to optimize the on/off state of the unit, and the Lambda-iteration method is adopted to solve the economic dispatch problem. In the iterative process, the solutions thatdo not satisfy all the constraints are adjusted by the correction method. Furthermore, different adjustment techniques such as conversion from cold start to hot start,decommitment of redundant unit, are adopted to avoid falling into local optimal solution and to keep the diversityof the feasible solutions. The proposed BGSO is tested onthe power system in the range of 10–140 generating units for a 24-h scheduling period and compared to quantum-inspired evolutionary algorithm (QEA), improved binaryparticle swarm optimization (IBPSO) and mixed integer programming (MIP). Simulated results distinctly show that BGSO is very competent in solving the UC problem incomparison to the previously reported algorithms.