AbstractTropical forests are being cleared at an alarming rate although our understanding of their ecology is limited. It is therefore essential to design restoration experiments that both further our basic knowledge of tropical ecology and inform management strategies to facilitate recovery of these ecosystems. Here we synthesize the results of research on tropical montane forest recovery in abandoned pasture in Costa Rica to address the following questions: (1) What factors limit tropical forest recovery in abandoned pasture? and (2) How can we use this information to design strategies to facilitate ecosystem recovery? Our results indicate that a number of factors impede tropical forest recovery in abandoned pasture land. The most important barriers are lack of dispersal of forest seeds and seedling competition with pasture grasses. High seed predation, low seed germination, lack of nutrients, high light intensity, and rabbit herbivory also affect recovery. Successful strategies to facilitate recovery in abandoned pastures must simultaneously overcome numerous obstacles. Our research shows that establishment of woody species, either native tree seedlings or early‐successional shrubs, can be successful in facilitating recovery, by enhancing seed dispersal and shading out pasture grasses. On the contrary, bird perching structures alone are not an effective strategy, because they only serve to enhance seed dispersal but do not reduce grass cover. Remnant pasture trees can serve as foci of natural recovery and may enhance growth of planted seedlings. Our results highlight the importance of: (1) understanding the basic biology of an ecosystem to design effective restoration strategies; (2) comparing results across a range of sites to determine which restoration strategies are most generally useful; and (3) considering where best to allocate efforts in large‐scale restoration projects.