This paper presents two stochastic search algorithms for optimizing the contract capacities of time-of-use (TOU) rate industrial customers. The proposed genetic algorithm (GA) and evolutionary algorithm (EA) are used to search the optimal contract capacities for the TOU rate customers to minimize the summation of demand and penalty charges. Because of the global search capabilities of the GA and EA, the proposed algorithms can optimize the solutions of the complicated, nondifferentiable, optimal contract capacity problem. The proposed algorithms have been tested on a three-section TOU rate industrial customer of Taipower system and are compared with the existing method. The test results demonstrate that the proposed algorithms can efficiently optimize the contract capacities for TOU rate customers.