Two largely independent bodies of scaling theory address the quantitative relationships between habitat area, species diversity and trophic interactions. Spatial theory within macroecology addresses how species richness scales with area in landscapes, while typically ignoring interspecific interactions. Complexity theory within community ecology addresses how trophic links scale with species richness in food webs, while typically ignoring spatial considerations. Recent studies suggest unifying these theories by demonstrating how spatial patterns influence food-web structure and vice versa. Here, we follow this suggestion by developing and empirically testing a more unified scaling theory. On the basis of power-law species-area relationships, we develop link-area and non-power-law link-species models that accurately predict how trophic links scale with area and species richness of microcosms, lakes and streams from community to metacommunity levels. In contrast to previous models that assume that species richness alone determines the number of trophic links, these models include the species' spatial distribution, and hence extend the domain of complexity theory to metacommunity scales. This generality and predictive success shows how complexity theory and spatial theory can be unified into a much more general theory addressing new domains of ecology.