The conversion of natural habitats to farmland is a major driver of the global extinction crisis [1 and 2]. Two strategies are promoted to mitigate the impacts of agricultural expansion on biodiversity: land sharing integrates wildlife-friendly habitats within farmland landscapes, and land sparing intensifies farming to allow the offset of natural reserves [3]. A key question is which strategy would protect the most phylogenetic diversity—the total evolutionary history shared across all species within a community [4]. Conserving phylogenetic diversity decreases the chance of losing unique phenotypic and ecological traits [5] and provides benefits for ecosystem function and stability [6 and 7]. Focusing on birds in the threatened Chocó-Andes hotspot of endemism [8], we tested the relative benefits of each strategy for retaining phylogenetic diversity in tropical cloud forest landscapes threatened by cattle pastures. Using landscape simulations, we find that land sharing would protect lower community-level phylogenetic diversity than land sparing and that with increasing distance from forest (from 500 to >1,500 m), land sharing is increasingly inferior to land sparing. Isolation from forest also leads to the loss of more evolutionarily distinct species from communities within land-sharing landscapes, which can be avoided with effective land sparing. Land-sharing policies that promote the integration of small-scale wildlife-friendly habitats might be of limited benefit without the simultaneous protection of larger blocks of natural habitat, which is most likely to be achieved via land-sparing measures.