The parasitoid Spalangia cameroni Perkins is a commercially available biocontrol agent of filth flies that is also used, especially through inundative releases, for the biological control of fruit flies. Adequate host selection behaviour is essential to the reproductive performance of parasitoid females, and innate and learned host species preferences play a key role in it. Therefore, in this study, the effect of natal host species [the medfly Ceratitis capitata (Wiedemann) or the house fly Musca domestica Linnaeus], either an innate or learned host, on later S. cameroni female host preference was assessed. In both innate and learned host preference experiments, most examined variables (realized fecundity, total progeny, induced mortality and sex ratio) did not differ significantly for the parasitoid line (parasitoids reared on medfly or house fly), but they did differ for the two attacked hosts, with the variables being higher in the house fly. In the learned host preference experiment, these variables did not differ significantly with experience (i.e., the host species on which parasitoids had performed previous parasitic activity). Superparasitism (the number of pupae with supernumerary eggs) differed significantly for the two attacked hosts and with experience, being higher in the house fly and in females experienced with medfly, respectively, but it did not influence the total progeny. The lower realized fecundity and induced mortality shown by experienced S. cameroni females compared to inexperienced females may be due to the fact that experienced females could be less likely to accept novel host types than naive females. These results indicate that rearing S. cameroni on medfly or house fly pupae seems to have little influence on the later performance of this parasitoid in its subsequent use against populations of both fly pests.