A key research problem faced by many ecologists is the extent to which the behavior of an ecological process within a landscape can be predicted from landscape structure and an understanding of fundamental mechanisms. For pollination systems at the scale of landscapes, the "holy grail" is: given a map of landscape composition and configuration, is it possible to predict where pollinators live, how they move to obtain resources, what the reproductive consequences are for the plants they visit, and whether the pollination system itself is sustainable? Research exploring the mechanisms underlying pollinator-plant interactions has productively focused on individuals, or within single populations at small spatial scales. In the past five years, ecologists worldwide have made unprecedented progress in identifying how landscape affects pollination service and pollinator communities. The purpose of this session is to explore the next step; that is, to identify and link mechanisms with landscape pattern that allow us to predict how landscape changes may affect pollinators and pollination services. The session develops a logical theme as follows. We use the "map" (landscape structure) as the organising concept. Three sections explore how we might use the map and our understanding of mechanism to predict: (1) pollinator abundance and viability; (2) pollinator movement; and (3) seed set, cross-pollination, and gene flow within landscapes. Because the goal is prediction from fundamental ecological principles, the symposium intercalates empirical studies with theory. We explicitly include contributors from various continents, recognizing that this will increase the diversity of perspectives. We also include speakers from disciplines beyond ecology to introduce new approaches and foster discussion of how to address our questions.