Aims: Ensuring sustainable agriculture in arid and semi-arid regions requires both structural rehabilitation of degraded soils and improved water use efficiency. This study evaluated the effects of maize-derived organic amendments maize straw (MS), maize compost (MC), and maize green biomass (MGB) on soil physical properties of clay-rich soils and developed a predictive model for soil rupture resistance using artificial neural networks (ANN). Methods: A three-year field experiment was conducted in Konya Province, Türkiye, using MS, MC, and MGB applied at 10, 20, and 40 Mg ha−1. Soil physical parameters, including macroaggregate stability (MAS), mean weight diameter (MWD), geometric mean diameter (GMD), bulk density (Pb), modulus of rupture (MR), penetration resistance (PR), and water use efficiency (WUE), were measured. Statistical analyses assessed treatment effects, and an ANN model was trained and validated to predict MR from soil physical indicators. Results: High-dose applications, particularly MC4, greatly improved soil physical quality, increasing aggregate stability 4.5-fold, reducing mechanical resistance by 65%, and lowering bulk density by 0.20 g cm−3. Compost also increased WUE by 223%. MGB4 showed comparable benefits, while straw effects were moderate. A clear dose-dependent trend was observed, and the ANN model accurately predicted MR (R2 = 0.87 training, 0.70 validation, 0.80 testing). Conclusions: Compost and green manure at elevated doses most effectively enhanced soil structure and water use efficiency in calcareous clay soils. The ANN approach provides a practical decision-support tool for evaluating soil mechanical behavior and supports sustainable soil management in arid and semi-arid regions.