The removal of invasive species is common in restoration projects, yet the long-term effects of pest management programs are seldom assessed. We present results of a long-term program to remove the invasive species Lupinus arboreus (lupin) from sand dunes in New Zealand. We evaluate the response of plant communities to lupin removal, by comparing total plant cover, the cover of non-native and native plant species, and species richness between sand dune sites where lupin removal has occurred, not occurred, and where lupin has never been present. Neither lupin presence nor removal had a significant impact on the foredune environment. Following removal, total and other non-native plant cover remained higher, and the cover of several native sand dune species remained lower compared with uninvaded sites in the deflation and backdune environments. These changes can be attributed to persistent effects associated with the invasion of lupin, but have also developed in response to lupin removal. The results of this study have implications for restoration projects in sand dunes. Pest management alone is unlikely to be sufficient to restore plant communities. Given the difficulties in restoring plant communities once an invasive species has established, managers should prioritize actions to prevent the spread of invasive species into uninvaded areas of sand dunes. Finally, the response to lupin invasion and removal differed between dune habitats. This highlights the importance of tailoring a pest management program to restoration goals by, for example, prioritizing areas in which the impacts of the invading species are greatest.