The optimal management of irrigation is a principal constraint facing agriculture worldwide. The potential number of combinations of irrigation schedules over a growing season is enormous when the variables involved are considered. Theoretically, on any given day during a cropping season, a decision can be made to irrigate or not. This associated with specific irrigation amounts and methods. The outcome of a decision and the nature of future décisions is afunction of the status of the crop at any given decision point. Thus, there are an extremely large number of potential irrigation schedules for any given crop season. Evolutionary algorithms are computer based optimization and search techniques that mimic natural selection to efficiently search very large solution spaces. They are based on the biological concepts of evolution through mechanisms such as selection, crossover and mutation. Cotton irrigation scheduling (timing and amount of water) can be approached as an adaptation problem. In this study a cotton simulation model (GOSSYM) was integrated with an evolutionary algorithm to provide irrigation recommendations to farmers. The results show that this approach Io irrigation scheduling is an improvement over the expert system currently associated with the model. This evolutionary algorithm consistently derived irrigation itineraries that resulted in higher economic return. More interesting, this new approach is also generic and it is applicable Io other cultural practices and to any cropping situation.