Habitat fragmentation, an important element of current global change, has profound repercussions on population and species extinction. Landscape fragmentation reduces individual movements between patches (i.e. dispersal) while such movements connecting patches enhance the persistence of metapopulations and metacommunities. Through the recognition of non‐random movements, dispersal has recently been recognized as a highly complex process. This complexity likely changes the predictions on the evolution of dispersal in spatially structured populations and communities. In this article, we emphasize the effects of fragmentation on the evolution of non‐random dispersal. Habitat fragmentation may shape local and global selective pressures acting on a large array of phenotypic traits known to covary with dispersal behaviors. On top of changes in dispersal propensity, habitat fragmentation could therefore modify dispersal syndromes (i.e. dispersers' phenotypic specializations). Habitat fragmentation often leads to spatial structuring of local conditions and consequently may lead to the evolution of different dispersal syndromes at the landscape scale. By neglecting impacts on dispersal syndromes, we might underestimate the impacts of fragmentation on a crucial biodiversity level for metapopulation and metacommunity functioning. We highlight a set of priorities for future empirical and theoretical work that together would provide the understanding of eco‐evolutionary dynamics of dispersal syndromes required for improving our ability to predict and manage spatially structured populations and communities.