Variation in plant functional traits has been related to variation in environmental conditions. In particular, the relationship between leaf traits and climate has received much attention. This paper presents a functional-trait-centred approach to identify potential impacts of climate and land use change on plant species assemblages. Using species atlas data, we modelled the relative frequencies of species with different leaf anatomies (LARF) as a function of observed climate and land use data on a regular spatial grid across Germany. Subsequently, we projected the geographical distribution of LARF with simulated climate and land use data for the late 21st century under two future scenarios. We used a conditional autoregressive regression model to account for spatially structured variation in LARF that remained unexplained by the environmental factors considered. We found a clear relationship between the climatic gradient of water availability and shifts in LARF: increasing water deficit was associated with a decreasing proportion of species with hygromorphic leaves in the composition and increasing proportions of species with scleromorphic and mesomorphic leaves. The variation in LARF due to land use was only small. Under future environmental scenarios the proportion of species with hygromorphic leaves was projected to decrease in all parts of Germany, whereas the proportions of species with sclero- and mesomorphic leaves were projected to increase on average. In particular, Germany's south-western and north-eastern areas were projected to experience functional change in LARF. Our study highlights the relationship between functional traits and plant species vulnerability to climate change. Our results suggest that the functional-trait-centred approach can provide a powerful additional modelling tool to estimate potential impacts of climate change on plant species assemblages.