Farm water management needs timely, spatially explicit guidance. We present an open-data framework that couples evapotranspiration (ET) with depth-resolved profile soil moisture (PSM) for field-scale irrigation scheduling. SEBAL ET (Landsat-8/9) is temporally harmonized with Sentinel-2 OPTRAM-ET to improve temporal continuity, while PSM at 0-20, 0-40, 0-60, 0-100 cm is mapped using Boruta-selected machine learning with ALE/SHAP interpretation. ET is harmonized to 30 m (-5-day) and drives a root-zone water-balance to trigger pixel-wise irrigation water requirement (IWR) using maximum allowable depletion. Across two wheat seasons, SEBAL agreed well with eddy-covariance observations (r ti 0.91; RMSE ti 0.51 mm d-1), while OPTRAM-ET performed competitively (r ti 0.89; RMSE ti 0.89 mm d-1). For PSM, the ensemble model achieved the best validation performance at 0-60 cm (r ti 0.86; RMSE ti 2.65%), with a depth-wise transition from spectral/LST controls near the surface to soil texture and terrain at depth; surface soil moisture strongly informed the root-zone up to-60 cm. In retrospective scheduling, the modeled IWR delayed the first three irrigation events by 10, 3, and 4 days, respectively, and estimated a seasonal irrigation requirement of 218 mm against 324 mm applied, indicating about 33% potential water saving. The framework combines temporally harmonized multi-sensor ET with interpretable, depth-resolved PSM to deliver temporally continuous, pixel-wise IWR at 30 m and-5-day resolution, with the OPTRAM pathway using a spatially diagnosed Priestley-Taylor (PT) coefficient. The approach offers a scalable pathway for precision irrigation management in semi-arid agroecosystems.