Aim Pollinators play an important role in ecosystem functioning, affecting also crop production. Their decline may hence lead to serious ecological and economic impacts, making it essential to understand the processes that drive pollinator shifts in space and time. Land-use changes are thought to be one of the most important drivers of pollinators’ loss, and there is increasing investment on pollinator-friendly landscape management. However, it is still unclear whether landscape history of a given region determines how pollinator communities respond to further landscape modification. Location The Netherlands. Methods Using geographically explicit historical landscape and pollinator data from the Netherlands, we evaluated how species richness changes of three important pollinator groups (bees, hoverflies and butterflies) are affected by landscape changes related to habitat composition, fragmentation and species spillover potential and whether such effects depend on the historical characteristics of the landscape. Results The effect of landscape changes varied between different pollinator groups. While bumblebee richness benefited from increases in edges between managed and natural systems, other bees benefited from increases in landscape heterogeneity and hoverfly richness was fairly resistant to land-use changes. We found that for the majority of the pollinators past landscape characteristics conditioned, the more recent pollinator richness changes. Landscapes that historically had more suitable habitat were more susceptible to display hoverfly declines (caused by drivers not considered in this study). Landscapes that historically had greater spillover potential were more likely to suffer butterfly richness declines and the bumblebee assemblages were more susceptible to the effects of fragmentation. Main conclusions The diversity of responses of the pollinator groups suggest that multispecies approaches that take group-specific responses to land-use change into account are highly valuable. These findings emphasize the limited value of a one-size-fits-all biodiversity conservation measure and highlight the importance of considering landscape history when planning biodiversity conservation actions.