This thesis gathers and analyses the local knowledge regarding ecosystem services in coffee producing regions of Costa Rica, Guatemala and Nicaragua, and compares this knowledge across a range of farming conditions. The extent to which coffee agroforestry systems provide ecosystem services depends on local context and management practices. There is paucity of information about how and why farmers manage their plantations in the way that they do and the local knowledge that underpins this. The present research compares local knowledge in coffee growing areas bordering key forest reserves in Nicaragua, Costa Rica and Guatemala. Knowledge was acquired from 99 coffee farmers in a stratified purposive sample, using established knowledge based systems methods. Farmers in all three countries had detailed knowledge about how trees affected ecosystem services such as soil formation, erosion control, provision of wildlife habitat and water conservation. A total of 135 tree species were mentioned by the farmers. Links between trees and biodiversity, pollination, biological pest control and micro-climate regulation were understood and species were classified according to their role in both provisioning and regulating services. Trees were said to produce ‘fresh’ shade that was suitable for coffee or ‘hot’ shade that was not suitable. This concept was widely used by farmers in the three countries; howerer, any coffee technician uses it. Fresh – hot dichotomy was explained in relation to leaf texture and size; foliage density, crown shape and root system attributes; as well as classification of trees regarding ecosystems functions such as water regulation or soil formation.Much of the local knowledge about how trees could improve provision of ecosystem services, however, was not practically applied because farmers were concerned that increasing levels of shade would decrease yields. A variety of tree species was maintained in coffee plantations at all sites but a few tree species were dominant. The degree of shade tolerated was the main difference across countries and this was strongly related to socio-economic factors such as the prevailing demand for fuel wood. Applicable knowledge across sites as well as the key factors that determine how knowledge was locally applied was identified.