Under increasing nutrient loading, shallow lakes may shift from a state of clear water dominated by submerged macrophytes to a turbid state dominated by phytoplankton or a shaded state dominated by floating macrophytes. How such regime shifts mediate the relationship between taxonomic and functional diversities (FD) and lake multifunctionality is poorly understood. We employed a detailed database describing a shallow lake over a 12-year period during which the lake has displayed all the three states (clear, turbid and shaded) to investigate how species richness, FD of fish and zooplankton, ecosystem multifunctionality and five individual ecosystem functions (nitrogen and phosphorus concentrations, standing fish biomass, algae production and light availability) differ among states. We also evaluated how the relationship between biodiversity (species richness and FD) and multifunctionality is affected by regime shifts. We showed that species richness and the FD of fish and zooplankton were highest during the clear state. The clear state also maintained the highest values of multifunctionality as well as standing fish biomass production, algae biomass and light availability, whereas the turbid and shaded states had higher nutrient concentrations. Functional diversity was the best predictor of multifunctionality. The relationship between FD and multifunctionality was strongly positive during the clear state, but such relationship became flatter after the shift to the turbid or shaded state. Our findings illustrate that focusing on functional traits may provide a more mechanistic understanding of how regime shifts affect biodiversity and the consequences for ecosystem functioning. Regime shifts towards a turbid or shaded state negatively affect the taxonomic diversity and FD of fish and zooplankton, which in turn impairs the multifunctionality of shallow lakes.