Power production plans are used by power companies to propose a feasible plan regarding the load requirements and system conditions of the following year, in which the arrangement of production material, such as the procurement of fuel required and delivery schedules, is clearly projected. Some of the factors involved in considering power production plans are common to those for short-term unit commitment, including system power balance, system security requirements, and unit characteristics; however, the limit on annual accumulation, which is not considered in short-term unit commitment, is not similar between the two. This study proposes an optimal mathematical model based on mixed-integer programming for long-term unit commitment to minimize the total system costs. To avoid the mixed-integer linear programming problems that increase the duration of computation because of numerous integer variables, the model proposed in this study applies the view of energy and power and integrates the similarities of load patterns and fuzzy logic to reduce the number of variables while adhering to all the constraints. The proposed algorithm can greatly improve the solution time compared with the complete UC model for the Taipower one week simulation case.