Abstract: Previous studies have demonstrated that higher nitrogen (N) and water availability affect both above-and below-ground communities, soil carbon and N pools, and microbial activity in semi-arid grasslands of Inner Mongolia. However, how soil phosphorus (P) and sulfur (S) pools, and related soil enzyme activities (as indicators of P and S cycles) respond to long-term N and water addition has still remained unclear. Since 2005, a field experiment with urea and water amendments has been conducted to examine their effects on total and available P and S concentrations and alkaline phosphomonoesterase (PME) and aryl-sulfatase (ArS) activities in three soil aggregate fractions: large macroaggregates (>2 mm), small macroaggregates (0.25-2 mm), and microaggregates (<0.25 mm) in an Inner Mongolia semi-arid grassland. Normalized to aggregate mass, microaggregates retained the highest total P and S concentrations. Both N and water additions increased the available P (by up to 84.5%) and the available S (by up to 150%) in the soil aggregate fractions. Soil acidification, as a result of the N addition, decreased both alkaline PME and ArS activities by up to 62.9% and 39.6%, respectively, while the water addition increased their activities. Our observations revealed that soil acidification (under the N addition) and elevated enzyme activity (under the water addition) played important roles in the levels of soil available P and S. The depression of P- and S-acquiring enzymes with soil acidification may decrease P and S availability, potentially impacting ecosystem processes and limiting the restoration of these grassland systems. The water addition was shown to be a more effective practice than the urea amendment for improving soil structure, supplying available P and S, and maintaining the sustainability of this semi-arid grassland.