Ecosystem restoration implies focusing on multiple trophic levels and ecosystem functioning, yet higher trophic levels, that is, animals, are less frequently targeted by restoration than plants. Habitat diversity, the spatial heterogeneity between and within habitat patches in a landscape, is a well-known driver of species diversity, and offers possible ways to increase species diversity at multiple trophic levels. We argue that habitat diversity is central in whole-ecosystem restoration as we review its importance, provide a practical definition for its components, and propose ways to target it in restoration. Restoration targeting habitat diversity is used commonly in aquatic ecosystems, mostly to increase the physical diversity of habitats, meant to provide more niches available to a higher number of animal species. To facilitate the uptake of habitat diversity in terrestrial ecosystem restoration, we distinguish between compositional and structural habitat diversity, because different animal groups will respond to different aspects of habitat diversity. We also propose four methods to increase habitat diversity: varying the starting conditions to obtain divergent successional pathways, emulating natural disturbances, establishing keystone structures, and applying ecosystem engineer species. We provide two case studies to illustrate how these components and methods can be incorporated in restoration. We conclude that targeting habitat diversity is a promising way to restore habitats for a multitude of species of animals and plants, and that it should become mainstream in restoration ecology and practice. We encourage the restoration community to consider compositional and structural habitat diversity and to specifically target habitat diversity in ecosystem restoration.

High‐quality information on predator–prey relationships is fundamental in understanding food webs, community assembly and ecosystem functioning. Recent analytical advances have made it possible to develop new trait‐based approaches to study trophic relationships and evaluate trait matching between predators and prey. Here, we develop a novel analytical approach based on generalized linear mixed‐effects models (GLMM) to test the importance of prey availability and to identify the set of prey traits that best explain the occurrence and number of prey in the predator's diet. We demonstrate that the approach by using an extensive dataset on prey availability, prey traits and gut content collected in all known populations of Vipera graeca, a little‐known, endangered snake of alpine grasslands in the Pindos Mountains of the Balkan Peninsula. We show that V. graeca is a unique, venomous snake specialized on bush‐crickets and grasshoppers (Orthoptera). Prey selection GLMMs showed that the ideal prey of V. graeca is abundant, large‐bodied, has poor escape abilities (flightless, slow‐moving and bad jumper) and prefers loose grasslands (as opposed to bare ground/rock or closed sward). Vipers restrict their feeding to periods of high Orthoptera abundance in the late summer and need to reach a certain body size to become able to catch large‐sized prey. Our analytical approach provides a framework for trait matching between predators and prey and unprecedented fine‐scale information on the importance of prey traits in prey selection by a specialist predator. The narrow trophic niche of V. graeca likely increases the vulnerability of this cold‐adapted snake to extinction. A free Plain Language Summary can be found within the Supporting Information of this article.