Background and aims – Species grouping in plant functional types (PFTs) according to Grime's CSR theory has been found useful in understanding ecosystem functioning and processes and in quantifying variation along environmental gradients. However, little is known about the relationship between herbaceous plant strategies and stress gradients in submediterranean forest ecosystems. Our aim was to assess variations in CSR strategies along three stress gradients (related to the amount of photosynthetically active radiation, soil water deficit, and soil total nitrogen content) in the herb layer of submediterranean forests. Material and methods – We collected floristic (species cover in the herb layer) and environmental data about topography, soil and light conditions in 48 sampling plots in a forest landscape of central Italy. We performed canonical redundancy analyses (RDAs) of species and PFT data sets, constrained by photosynthetically active radiation, soil water deficit and soil total nitrogen content. Cover trends of PFTs along the environmental gradients were graphically represented. Key results – All the environmental factors considered did affect ecological strategies in the herb layer, but the variation of both species and PFT data sets were best explained by photosynthetically active radiation. Competitors were favoured by low/intermediate stress levels, with regard to soil water and light intensity. Competitive stress-tolerant ruderals peaked in moderate light stress intensity and in well-watered soils. Stress-tolerant competitors peaked in conditions of strong light and moderate dryness. Conclusions – Spatial variations in environmental conditions and in the related stress gradients determine predictable variations in the functional composition of the herb layer. Therefore, it may be possible to use Grimes's PFTs for assessing environmental constraints and predicting whether a species with a specific strategy is likely to be positively or negatively affected by changes in environmental stress due, for instance, to climate change.