Species-rich grasslands contribute significantly to conserving environmental quality in Europe but have suffered for decades due to area reduction and degradation from nutrient addition of agricultural or pollution origins. Studying the effects of grassland fertilization can supply useful information on their degradation patterns and restoration possibilities.On a valley meadow in the Italian Alps fertilized with farmyard manure until 1976, a fertilization trial with 4 replicates and 27 treatments from the factorial combination of three levels of nitrogen (N), phosphorus (P) and potassium (K) was established in 1977 and surveyed until 2003 for botanical composition on 18 m2 plots. The change of species composition and species richness (SR) occurring over the surveying period was analyzed.Six types of vegetation successions were identified, mostly with homogeneous nutrient combinations and only marginal influence by the amount of nutrients added. In all successions, the vegetation change was significant compared to the pre-experiment situation and more rapid but less persistent for abundance than for presence-absence. From about 33.6 species per plot surveyed in 1976, SR had the highest increase in the succession with no-fertilization or only-P or only-K addition (40.2 species in the period 1991–2003). The N + P + K (27.2 species) and N + P (26.3 species) successions had the strongest negative impact on SR. In the other successions (N + K, P + K and only N) SR increased or decreased little. No soil acidification and no negative effect on SR from the acidifying N-fertilizer ammonium nitrate was observed thanks to the buffering power of the Ca‑carbonate soil content.In the two successions with the highest SR variation the SR changed rapidly in the first period but fluctuated later around a constant value. In the succession with significant but little SR variation, this trend was not obvious due to the considerable interannual SR fluctuation. In all successions, the SR change was the result of a balance between logarithmic trends of species loss and gain. Species loss was probably caused by stochastic extinction of low-abundance species, but in NP, it was also due to the establishment of the aggressive-growth-habit Festuca rubra. For species gain, the suitability of incoming species to new fertility conditions was an important factor.Trial results showed that in calcareous, permeable soils, the species richness of central European grassland is especially endangered by the combined addition of N and P (with or without K), more than the only-N addition. However, they also showed that the restoration of grassland that had been degraded due to high NPK inputs from agriculture and NP enrichment from water and atmosphere pollution is possible with the cessation of fertilization and without seed addition if the surrounding landscape is species-rich and plants with aggressive growth habits are not established. Management for grassland biodiversity conservation should avoid the utilization of soil-acidifying fertilizers, reducing the P input beside the N input and limiting the introduction and spreading of aggressive grass species, such as Festuca rubra.