Enhancing nitrogen use efficiency (NUE) and crop water productivity (WPc) in drought-resistant wheat under limited irrigation is essential to mitigate water scarcity in the North China Plain (NCP). However, the microbial-mediated mechanisms under different irrigation regimes remain unclear. A 2-year field experiment (2018–2020) was conducted with two winter wheat cultivars, namely-Jimai 22 (J22) and Shinong 086 (S086), under three irrigation treatments: no irrigation after the overwintering stage (W0), single irrigation at the jointing stage (W1, 75 mm), and irrigation at the jointing and the booting stages (W2, 75 mm each). We determined wheat grain yield, crop evapotranspiration (ETc), and grain nitrogen accumulation at maturity to calculate WPc and NUE. The relationships among WPc, NUE, and core OTU and between enzyme activities and volumetric water content (VWC) were revealed for both wheat varieties. We also examined the interactions among irrigation practices, wheat varieties, and microbial dynamics by integrating soil microbial community analysis, co-occurrence network modeling, and enzyme activity assays. Data were analyzed using analysis of variance (ANOVA) and t-tests, with statistical significance set at P < 0.05. Compared to W0, grain yield increased by 121.29 % and 163.82 % (P < 0.05) in W1 and W2, respectively. Under W1 conditions, the NUE of S086 was 6.83 % higher than that of J22, whereas the WPc of S086 decreased significantly by 6.92 % compared to that of J22 (P < 0.05). Both irrigation regime and wheat variety significantly influenced the composition and diversity of bacterial and fungal communities, with notable interactive effects (P < 0.05). Co-occurrence network analysis revealed a higher complexity (highest average cluster coefficient; edge count=213) in fungal networks of S086 compared to those of J22 (edge count=193) under W1. Keystone genera such as like Trichoderma were linked to enhanced NUE via nitrogen mineralization. Our integrated analysis of soil microbial communities, co-occurrence networks, and enzyme activities revealed that fungal communities are more sensitive to irrigation and varietal differences than bacterial communities. These findings emphasize that by linking microbial dynamics to resource use efficiency, we can propose strategies for manipulating rhizosphere microbiomes to enhance drought adaptation, offering significant potential for improving resilience in water-limited agroecosystems.