The global trend for industries to comprehensively improve the utilization of resources has attracted great attention from many researchers. Industry managers are increasingly optimizing their production strategies under time-of-use (TOU) electricity tariffs to reduce energy consumption since it accounts for the majority of total production costs. This paper investigates the single-machine scheduling problem with release dates under TOU tariffs. The key issue is to assign a group of jobs with different release dates to a single machine to minimize the total electricity cost. To solve the problem, a mixed-integer linear programming (MILP) model based on time-indexed formulation (TI-MILP) is proposed. By preliminary experiments, we find TI-MILP model is not efficient enough. Therefore, to more efficiently solve the problem, a period-based MILP (P-MILP) model is developed. Finally, extensive experimental results demonstrate that i) the proposed models can save about 25% on total electricity cost compared with the existing empirical scheduling method, and ii) P-MILP model outperforms TI-MILP model in terms of computational efficiency and problem scale.