Host plants affect development, survival, and reproduction of phytophagous insects. In the case of holometabolous species, whose larvae have little mobility to find a host plant, the ability of females to discriminate hosts on the basis of their nutritional quality may be an important factor determining insect performance. The preference—performance correlation hypothesis states that females will choose to lay their eggs on host plants that provide the best offspring performance. The effects of three cultivated and two wild brassicas (Brassicales: Brassicaceae) on the biology of the diamondback moth, Plutella xylostella L. (Lepidoptera: Plutellidae), an important pest of brassicas, were investigated. Based on these data, the preference-performance correlation hypothesis was tested. The results allowed the discussion of the possible role of wild brassicas on population dynamics of the pest. The life table parameters net reproduction rate and intrinsic rate of increase were used as indicatives of insect performance because they provide a detailed description of the survivorship, development, and reproduction of a population. Development, survival, and reproduction were affected by the cultivated and wild brassicas. Both net reproduction rate and intrinsic rate of increase were lower in individuals fed on wild brassicas, which indicates that brassicas are not nutritionally suitable for P. xylostella. Nevertheless, females showed no oviposition preference among host plants. The results showed that host plant quality might not be the only factor determining host selection by female P. xylostella. Results also suggest that wild brassicas may serve as a refuge for P. xylostella, favoring pest survival when crops are disturbed by insecticide application, irrigation, or ploughing.