Theoretical studies have predicted that reproductive interference must exclude either of the interacting species, but no testing of this prediction has ever been reported for natural populations. This study surveyed the distribution patterns of herbaceous Veronica plants, including one native and three alien species, to test whether the invasion of the alien species exerting reproductive interference excluded the native species. Results showed that the native species was repeatedly excluded from islands where an alien species invaded, exerting reproductive interference, and that other alien species had no significant effect on the native population survival. This survey also demonstrated that the native species altered its habitat from the ground to stone walls on the mainland where the alien species had been predominant. In the mainland populations, the fruit morphology differed from that of the islands, and the morphology in the mainland population seemed suitable for seed dispersion by ants at a stone wall habitat. We also surveyed the genetic differentiation among populations, the results of which did not support the native species genetically differentiated between mainland and island populations before the alien species invasion.These results strongly suggest that the reproductive interference excluded the recipient species at the population level and facilitated the habitat change. Additionally, results indicated that a series of field surveys of islands close to the mainland can be a powerful tool to test the ecological importance of reproductive interference.