A method to optimize the on-time performance of containerships is proposed. Two optimization criteria are considered: the maximization of the percentage of on-time arrivals and the minimization of the magnitude of arrival delays. In both cases, the problem is posed as a ship scheduling problem in which the randomness of cargo volumes, port productivity, and sea conditions is considered. Fuel oil costs corresponding to different levels of service are investigated. A solution method involving discrete simulation and nonlinear programming is developed. A numerical example is presented to illustrate the method.