Coffee is widely grown across Central America at altitudes between 600 and 2500 m, mostly in association with trees that provide shade and other services. Research on coffee agroforestry systems has identified many environmental factors, management strategies and plant characteristics that affect growth, yield and environmental impact of the system. Much of this literature only presents qualitative estimates of the importance of the different growth determining factors, or highly site-specific estimates. Quantitative information is required to allow statistical analysis or the construction of process-based models of the system. Here, we review the available quantitative information for the latter purpose, with emphasis on the data needs for modelling agroforestry systems common in Central America. Process-based models require environmental data-weather, soil-and data on the physiological characteristics of the coffee plants and trees. Our review showed that the current literature is insufficient to allow full parameterisation of a process-based model for any coffee-tree combination. Information on weather, coffee and trees is highly limited, but soil information seems more adequate. A regional network of replicated multi-factorial experiments, focusing on the interactive effects of different environmental factors, may help address the main knowledge gaps.