The combination of social and ecological factors determines people's vulnerability to climate change (CC) and options for climate smart agriculture. This study uses the tree cover transition curve framework (TTC that links reforestation stages of landscapes to a preceding deforestation process), to understand how adaptive capacity and perception on CC vary as agricultural systems develop and tree resources decrease. We selected 3 sites in Nicaragua (Siuna, Waslala and La Dalia) to represent tree cover decline phases (from early degradation: Siuna, towards sites with low tree cover: La Dalia) and Hojancha-Costa Rica to represent recovery phase. We characterized 140 households per site with a household survey (www.ccafs.cgiar.org). Different sets of 45 indicators (e.g. livelihood capitals, food security, land use, farm size, and CC induced changes in farm management) were used in 4 cluster analyses to explore how farm types (land use allocation, CC perceptions, adaptive capacity and available tree resources) vary along the TTC. At early degradation and recovery phase, extensive cattle ranching and basic grains dominate, there are few progressive CC-induced changes in farm management (but high impacts of extreme events in Siuna), good adaptive capacity and good food security. Intermediate phase is similar to early degradation in land use and CC perception, but with the lowest adaptive capacity. In agricultural landscapes where basic grains smallholders agriculture predominate, most farmers have changed farm practices due to progressive CC, have low to medium adaptive capacity and high food insecurity. Early and intermediate degradation landscapes offer opportunities for mitigation, fostering sustainable intensification; improving coping capacity of farmers there might offer triple wins. In landscapes with high pressure on land (La Dalia) options for mitigation are limited; efforts for improving adaptive capacity and alleviation of food insecurity are perhaps the only i