Research to understand and manage ecosystems to supply services has recently spurred a functional view of their biodiversity. In particular, approaches based on functional traits rather than species diversity are increasingly used to reflect interactions between organisms and their environment. These approaches bring a functional perspective to the study of community structure responses to disturbances and resources, and of their effects on ecosystem functioning and services. From an academic perspective, we propose a conceptual framework based on species functional traits to better infer how grassland management practices (fertilization, defoliation regime) along with abiotic factors influence plant, animal and microbial community composition and a range of services in grassland ecosystems. The core of the framework relies on combinations of plant functional traits and associated microbial features that specifically respond to environmental and management factors and influence ecosystem services. To overcome stakeholders’ difficulty in applying the concept of functional traits, we propose an operational approach implying the mapping of plant communities distributed into five plant functional types (PFTs). The approach was used for fields in grassland-based livestock farms from two French grassland networks. We evaluated its ability to predict a range of services including forage provision and non-market services according to environmental and management drivers. PFT-based plant community composition predicted forage services reasonably well but responded weakly to environmental gradients. To cope with the observed limitations of current predictive approaches, we suggest including soil microbial functional types and adaptive management rather than using a prescriptive scheme.