Agricultural system models are useful tools to synthesize field experimental data and to extrapolate results to longer periods of weather and to other cropping systems. The objectives of this study were: (i) to quantify the effects of crop management practices and tillage on soil water and spring wheat (Triticum aestivum L.) production in a continuous spring wheat system using the RZWQM2 model (coupled with CERES-Wheat) under a dryland condition, and (ii) to extend the RZWQM2 model results to longer term weather conditions and propose alternate cropping systems and management practices. Measured soil water content, yield, and total aboveground biomass under different tillage and plant management practices were used to calibrate and evaluate the RZWQM2 model. The model showed no impacts of tillage but late planting greatly reduced grain yield and biomass, in agreement with observed differences among treatments. The hydrologic analysis under long-term climate variability showed a large water deficit (32.3 cm) for spring wheat. Fallowing the cropland every other year conserved 4.2 cm of water for the following wheat year, of which only 1.7 cm water was taken up by wheat, resulting in a yield increase of 249 kg ha⁻¹ (13.7%); however, the annualized mean yield decreased 782 kg ha⁻¹(43.1%) due to 1 yr of fallow. Other long-term simulations showed that optimal planting dates ranged from 1 March to 10 April and the seeding rates with optimum economic return were 3.71 and 3.95 × 10⁶ seeds ha⁻¹ for conventional and ecological management treatments, respectively.