Understanding why certain plant communities are vulnerable to alien invasive species is essential to predicting and controlling invasion in a changing environment. Darwin’s naturalization hypothesis suggests that non-native species should be more successful in communities where their close relatives are absent. Empirical tests of this hypothesis, however, have produced mixed results. Using plot-level data from natural forests along elevational transects covering strong environmental gradients, we examined whether the invasion of the globally invasive species Ageratina adenophora can be explained by environmental filtering and/or competition from closely related species linked to environmental gradients. Abundant precipitation, warm temperatures, open canopies, and postfire environments facilitated A. adenophora invasion, whereas resident taxonomic richness suppressed its invasion. Importantly, we found that invader-resident relatedness had a strong negative effect on invader cover under resource scarcity conditions (e.g., low water availability), but not under nonresource environmental stress conditions (e.g., low temperature). Our findings help reconcile the varied applicability of Darwin’s naturalization hypothesis to biological invasions in a changing world.