In arid regions, the ecological restoration of limestone tailings requires sustainable strategies, yet the synergistic effects of substrate optimization and native plant selection remain poorly understood. In this study, we systematically evaluated substrate amendments and native species for rehabilitating limestone tailings in Northern China's arid zone using a controlled pot experiment. An orthogonal L9(34) experimental design was employed to test three factors: the soil-to-tailings ratio (1:2, 1:1, and 2:1), moisture level (30%, 45%, and 60% of field capacity), and nitrogen addition (0, 5, and 10 g N m-2). Five native grass species (Pennisetum centrasiaticum, Setaria viridis, Leymus chinensis, Achnatherum splendens, and Eleusine indica) were grown under these treatment conditions, and plant biomass and key soil nutrient variables were measured. Stepwise regression, structural equation modeling, and principal component analysis were applied to assess plant growth responses and soil nutrient dynamics. The results indicated that a 2:1 soil-to-tailings substrate maintained at 60% moisture content maximized biomass production across all species. Soil total potassium consistently correlated positively with biomass (Standardized beta: 0.397-0.603), whereas available potassium showed a negative relationship (Standardized beta: -0.825--0.391). Nutrient dynamics ultimately governed biomass accumulation, accounting for 57.8-84.2% of the biomass variation. P. centrasiaticum ranked as the most effective species, followed by S. viridis, L. chinensis, A. splendens, and E. indica. We concluded that successful restoration under these experimental conditions hinged on key factors: using a 2:1 soil-to-tailings substrate, maintaining 60% soil moisture, and strategically combining deep-rooted P. centrasiaticum with shallow-rooted S. viridis to exploit complementary resource use. This work provides fundamental data and a conceptual framework for rehabilitating arid limestone tailings in similar ecological settings, based on controlled experimental evidence.