Pests and diseases are the main yield-reducing factors in simplified agricultural systems. Their role in complex, diverse, agricultural systems may, however, be less apparent because of the many interactions in which they are involved. However, it is essential to understand the relationships between pests and diseases, on the one hand, and ecosystem services (including crop production), on the other, to develop sustainable agroecosystems. Our study aims to illustrate these complex relationships based on the example of coffee agroecosystems in Costa Rica. We analysed a dataset consisting of 107 coffee plots characterized for their topoclimates, soils, coffee plant production characteristics, cropping practices, and pest and disease injuries. Meta-variables were created through cluster analyses to account for these different broad attributes of coffee-based agroecosystems. In particular, coffee injury profiles were determined on the basis of injury levels incurred by pathogens, nematodes, and insects over the course of one growing season. We used correspondence analysis to assess the levels of linkage between injury profiles and other agroecosystem meta-variables. Indicators of biodiversity based on shade diversity and of attainable yield were incorporated in the analysis as additional variables. Four groups of coffee-based agrosystems were identified, ranging from extensive (low-input, perennial polyculture) to intensive (unshaded high-input monoculture). Each group of coffee-based agroecosystem corresponds to varying levels of pest and disease injuries, crop yield, and ecosystem service provision, excluding coffee production. In each group of coffee-based agrosystem, we discussed the drivers of coffee production and explored potential avenues to improve sustainability based on ecosystem services provision. We highlighted that the physical characteristics of the environment, topoclimate and soil characteristics, are the main drivers of injury profiles and of resultin