This paper proposes a two-stage optimization method to solve the unit commitment of a large-scale system. First, the improved priority list method is used to quickly arrange the start-stop of the unit. Three parameters of the adjustable generator unit coal consumption, unit maximum output, unit startup cost is introduced to comprehensively reflect the economics of the generator; different from the traditional PL method, which firstly arranges the unit to start and stop to meet the rotation reserve, and then adjusts the unit state to meet the minimum start-stop time. This paper considers these two constraints at the same time to better achieve the priority to start the unit with lower operating cost. The improved particle swarm optimization algorithm is used to optimize the unit's economic dispatch. Tthis paper proposes a power balance adjustment strategy considering the unit's climbing rate, which ensures the diversity of particles while avoiding rapid convergence to the local optimal solution. The proposed method is applied to 10 units, 20 units, 40 units, 60 units and 80 units, and 100 units and 24 periods of time systems to be tested, and compared with other methods. The numerical results show that the proposed two-stage optimization method is fast, and its convergence is good and can effectively solve large-scale system unit combination problems.