The aim of this study was to investigate how a screening program interacts with the dynamic natural history of cervical neoplasia in order to learn how screening for cervical cancer should be carried out most cost-effectively. For the purpose of optimization, the fundamental quantities of the natural history are the shape of the curve of age-specific prevalence rates of cancer in situ, and the duration of this stage before progression to invasive cancer. Since these quantities seem similar in different populations, the results probably can be generalized. Simulation studies revealed that optimal screening implies a nonperiodic schedule, and that this schedule is dependent not only on the number of screenings, but also on the efficiency of the screening rounds. With lower efficiency, the schedule should be concentrated into a shorter period. The effect of screening is also especially sensitive to the choice of age at the first and--to a smaller extent--last screening. The results concerning three different objectives also were compared, namely reduction of the cumulative number of invasive cancers, reduction of deaths due to cervical cancer, and reduction of years-of-life lost. The choice of objective also plays a role in the design of an optimal screening program. The theoretical approach in this analysis easily can be developed and accommodated to more complex individual situations such as prior screening histories and differing screening schedules in high- and low-risk women.