Understanding how orchard water use and root-water uptake vary with stand age under contrasting moisture conditions is essential for developing precision water management in cold-arid regions. Sap flow monitoring and xylem stable isotope analysis were integrated to compare the water-use strategies of 5- and 10-year-old Malus asiatica stands in Inner Mongolia, China. Sap-flux density, stand water use (Q), canopy stomatal conductance (gc), and the canopy decoupling coefficient (Ω) were analyzed, and the contributions of soil water from different depths (0–80 cm) were quantified using δ18O. The results showed clear stand age- and moisture-dependent differences in both canopy water use and water uptake. Across the entire experimental period, total Q in the 10-year-old stand was 2.27 times that of the 5-year-old stand, and both Q and gc differed significantly between the two stands, with the greatest difference occurring during the wet period. During dry periods, the 5-year-old stand showed an earlier sap-flow peak followed by a rapid decline and no obvious midday plateau. The 10-year-old stand maintained a later peak and a broader plateau, indicating stronger capacity to sustain transpiration around midday. During dry periods, the decoupling Ω remained low in both stands being 0.04 in the 5-year-old stand and 0.09 in the 10-year-old stand. But decoupling increased markedly during the wet period, 0.07 and 0.20, respectively, with a steeper Ω–vapor pressure deficit (VPD) slope in the 10-year-old stand. This indicates greater plasticity in canopy-atmosphere coupling. Correlation analyses further indicated stronger soil-water limitation in the 5-year-old stand, whereas water use in the 10-year-old stand remained more responsive to atmospheric drivers. The strongest divergence in water uptake occurred during the wet period, when the 10-year-old stand relied mainly on shallow soil water, whereas the 5-year-old stand maintained a greater contribution from deep soil water. During dry periods, the contribution of deep soil water increased slightly in both stands, but shallow soil water still remained an important source. Overall, the 5-year-old stand exhibited a relatively conservative strategy characterized by lower Q, lower gc, and stronger dependence on deep soil water, whereas the 10-year-old stand showed a more opportunistic strategy that combined higher Q and higher gc with stronger exploitation of shallow soil water when it was available. Because the orchard was rainfed and SMSI remained below 0.4 during most of the growing season, these age-specific differences provide a basis for irrigation scheduling: moderate supplemental irrigation is needed in young stands to relieve early-season soil-water limitation and reduce dependence on deep soil water, whereas irrigation in mature stands should focus on rapidly maintaining shallow root-zone water availability during periods of high transpiration demand.