The increase in the proportion of renewable energy has led to a higher degree of grid interconnection, an increase in the number of random variables, and a greater number of generating units in unit commitment (UC) computation. Solving algorithms based on branch-and-bound methods can no longer meet the computational requirements for UC. This paper proposes a dynamic priority list approach that reveals the correlation between unit performance and constraint activation, and determine must-on and must-off units, thereby reducing the integer variables associated with the unit state and eliminating the majority of non-optimal iterative schemes to narrow the computation scale of the UC problem. A comparative case study is conducted on IEEE 118-bus system and ACTIVSg10k system. The results verify that the proposed method improves the accuracy of unit ranking, effectively reduces the scale of the UC problem, and reduces the computational time of problem optimization.