To study the contribution of moisture from different straw-treated and irrigated soil layers to the water consumption of winter wheat in dry farming, a 2-year straw treatment and regulated deficit irrigation experiment was implemented. The field experiment was carried out with 0% (S0), 1% (S1), and 2% (S2) straw returning amounts, and 75 mm (V3), 60 mm (V2), and 45 mm (V1) irrigation volumes. This experiment involved nine treatments, used to quantitatively analyze the ratio and variation of soil water use from different soil layers via the direct contrast method (DCM) and the multiple linear mixed model (MLMM). The results show the following: (1) The distribution of precipitation isotope compositions displayed a repeated trend of first decreasing and then increasing during the study period. Regression analysis showed that the local meteoric water line (LMWL): δD = 6.37δ18O − 3.77 (R2 = 0.832). (2) With increasing soil depth, the δ18O value decreased gradually, and the maximum δ18O value of the soil water within each growth period was distributed at 10 cm. (3) Under the same irrigation amount, δ 18O increased with increasing straw return at 0–20 cm and decreased with increasing straw return at 20–80 cm. (4) The comparison results of the DCM and MLMM were consistent. During the jointing and flowering stages, 0–30 cm soil water was the main source of water for winter wheat. The contribution of soil water below 30 cm had a decreasing trend from the jointing stage to the flowering stage. The average contribution rates of the 0–30 cm soil layer during the jointing and flowering stages were 23.07% and 23.15%, respectively. These findings have important implications for studying the soil water cycle in the context of farming.