Smallholder farms that transition to organic and biodiverse production are increasingly recognized as strongholds of agrobiodiversity, with emerging work identifying important outcomes such as enhancing crop portfolios, mitigating extreme climate events and contributing to farmer well-being. Yet the emergent herbaceous communities in these organic systems remain understudied, with the functional diversity and management of this stratum relatively unknown. This study identifies the taxonomic and functional diversity of the herbaceous community in organic coffee agroforestry systems, and describes the extent of this diversity with farm, and farmer, attributes. We measured leaf-level functional traits (e.g. specific leaf area) of the herbaceous community to derive functional diversity indices and collected localized environmental conditions on 15 organic coffee farms in Central Valley, Costa Rica. We also conducted semi-structured interviews with nine farmers to construct mental models on herbaceous community management using a cognitive mapping approach. In total, 38 species from 20 taxonomic families were present in these organic coffee systems. The herbaceous communities were functionally diverse; however, functional evenness increased with canopy openness, suggesting that farms adopting agroforestry tend to have a more functionally diverse herbaceous stratum. Farmer perception of plant traits in the herbaceous community was differentiated into competitive (weeds) or neutral/positive effects. These perceptions aligned with well-established functional trait trade-offs. The mental models representing farmer decision-making processes were highly variable, with a nearly 30% increase in cognitive map density from the simplest map to the most complex; this complexity in mental models was a key explanatory variable in the level of functional diversity of the herbaceous community. Organic management practices that support agroforestry practices also, in turn, promote a func