Background and aims <p>Nitrogen (N) and phosphorus (P) are essential for ecosystem productivity and nutrient cycling, yet their balance is increasingly altered by human-driven N enrichment. However, how different N forms affect plant nutrient status and stoichiometry in karst grasslands remains unclear.</p> Methods <p>We established a controlled pot experiment with two typical karst soils from southwestern China (yellow and limestone soils), applying four N forms (Ca(NO₃)₂, NH₄Cl, NH₄NO₃, and urea) at two application rates. Biomass, N and P concentrations, and N:P mass ratio were measured in perennial ryegrass and white clover, as well as at the community level.</p> Results <p>N application significantly increased plant shoot biomass, with ammonium-based forms (urea and NH₄Cl) producing the strongest effects and yellow soil exhibiting higher sensitivity. Perennial ryegrass biomass increased under all N treatments, while white clover biomass declined under high ammonium input. However, N addition generally reduced plant N and P concentrations in tissues due to dilution effects, which were most pronounced under NH₄Cl treatments. Community-level N:P mass ratio remained relatively stable, primarily influenced by the proportion of white clover, indicating a buffering effect at the community scale.</p> Conclusion <p>Ammonium-based N inputs produced the strongest biomass gains but also the largest declines in shoot N and P, consistent with growth dilution and reduced P availability under ammonium-induced acidification. Responses were consistently stronger in yellow soil than in limestone soil. At the community level, the N:P mass ratio was largely buffered by white clover, decreasing only under high NH₄Cl where clover abundance declined. These findings highlight that N form is a key, often overlooked driver of plant stoichiometry in karst grasslands and support N-form-aware fertilization with P co-management to sustain productivity while minimizing nutrient imbalance.</p>