1. Understanding the relationship between biodiversity and ecosystem functioning (BEF) is a central topic in ecology. Multi-trait-based functional diversity has been proposed to improve mechanistic understanding of the BEF relationship; however, how trait-based functional diversity affects ecosystem functioning and processes has rarely been addressed in aquatic ecosystems. 2. Here, we examined the causal relationships between three phytoplankton functional diversity indices (FAD2, functional diversity based on dendrograms [FDc], FRic) and Shannon diversity index versus resource use efficiency for nitrogen (RUEN), phosphorus (RUEP) and silicate (RUESi), with monthly long-term datasets from the marine (Western English Channel, 2000-2014) and freshwater (Lake Kasumigaura, 1984-2012) ecosystems. 3. We employed Convergent Cross Mapping (CCM), a novel method developed for identifying causality for nonlinear dynamical systems; this is in contrast to linear approaches that cannot distinguish causality from correlation. CCM found that FDc is the most robust functional diversity index among the selected functional diversity indices (FAD2, FDc, FRic) in predicting phytoplankton resource use efficiency and exhibited a much stronger causal effect than the Shannon index. 4. Furthermore, scenario exploration analysis indicates that most causal effects from phytoplankton diversity indices on resource use efficiency (RUEN, RUEP and RUESi) are on average positive, and FDc exhibited the most consistent positive causal effects on phytoplankton resource efficiency in both marine and freshwater ecosystems. Thus, increasing FDc can enhance phytoplankton resource use efficiency in aquatic ecosystems. 5. Synthesis. Our results show significant causal effects of functional diversity on phytoplankton resource use efficiency in both marine and freshwater ecosystems. Among all selected functional diversity indices, functional diversity based on dendrogram is the most robust functional diversity index in promoting phytoplankton resource efficiency. Our study provides empirical evidences in natural aquatic systems that trait-based functional diversity represents better species niche partitioning than the Shannon index and thereafter enhances resource use efficiency. This finding can improve our understanding on trophic transfer and nutrient cycling in aquatic ecosystems.