In complex coffee-based agroforestry systems, quantifying the impact of shade trees on coffee disease regulation and coffee yield is crucial for improving these systems and designing more sustainable ones. To this end, we analyzed interactions amongst shade trees, coffee plants (cv. Catimor), the coffee foliar disease complex and soil characteristics. We studied systems characterized by 40 variables measured in 60 plots located on three farms (monitored for 2 years) in Nicaragua. These variables characterized six system components grouped in six statistical blocks: shade trees (shade percentage and species abundancy), soil characteristics (fertility), foliar diseases, coffee plant characteristics (age and size), coffee growth and yield. We used partial least square path modelling (PLS-PM), i.e. a structural equation modelling approach used to understand and quantify interactions between the six blocks. Shade trees (mostly the associated shade percentage) had direct positive effects on foliar disease severity and incidence and soil quality, while having negative effects on coffee growth and yield. Soil characteristics (carbon, nitrogen, litter index, water infiltration potential) were negatively correlated with foliar diseases. An excessive shade percentage then had an indirect negative effect on coffee growth and yield due to the increased prevalence of foliar diseases. Finding the optimal shade cover can help reduce foliar diseases and enhance coffee berry production. The 'dose effect' of shade cover must also be considered because excessive shade, as well as lack of shade, have negative impacts on coffee growth and yield. Overall, effective shade management requires an analysis of trade-offs between soil quality, disease regulation and yield gains. In conclusion, PLS-PM turned out to be a good tool for studying agroecosystem networks and enabled us to put forward some foliar disease management and coffee yield enhancement guidelines.