Climate change is projected to result in shifts in the geographical distribution of ecosystems and species. Biogeography studies that predict the changes in life zone distribution allow evaluating the potential impacts of climate change on ecosystems. However, impact studies must be associated with an estimation of adaptive capacity in order to assess ecosystem vulnerability. Plant migration is a potential response from ecosystems for adapting to climate change. The migration capacity depends on plant types and on the fragmentation of landscape through which species will disperse. However, the development of plant migration models and the collection of data on migration rates remain a challenge. We developed a conceptual model to represent the impact of climate change on ecosystem distribution considering the capacity of organisms to migrate in a fragmented landscape. This model is implemented with a cellular automata, in which each cell is a landscape unit, characterized by its membership to life zone classes, its richness, and its state (natural vs. anthropic). Life zones are used as a proxy for ecosystem types and are assumed to be composed of different groups of plant functional types with different hypothetical migration capacities. The model is applied to the real landscape of the Central American region. Current landscape fragmentation is assessed using land use maps. Landscape scenarios represent either further fragmentation or connectivity enhancement based on the current distribution of the Protected Areas and Biological Corridors. A climate change scenario is applied to the simulated landscape to evaluate ecosystem shifts, under different landscape scenarios. Results show that well-designed conservation plans enhancing connectivity could increase ecosystem resilience to climate change depending on the design of the connectivity network. (Texte intégral)