Abstract We describe the development and use of a simulation model to predict the patch residence time and rate of parasitism of the wasp Aphelinus asychis Walker (Hymenoptera: Aphelinidae), a solitary parasitoid of the pea aphid. We use the model to test the hypothesis that a wasp's decision to leave a patch or to accept or reject a host is based on information about previously encountered hosts that are stored in a sliding “memory window”. The model incorporates the foraging behavior of individual wasps, including a simplified learning process of the proportions of parasitized and unparasitized hosts among all hosts encountered. Optimal foraging theory predicts that an animal should spend more time in high-quality than low-quality patches. The simulation results are consistent with this prediction as well as with empirical laboratory data. We propose that searching females of A. asychis store information about a changing environment in five units of a sliding “memory window”.