Microbial inoculants are increasingly used in conservation agriculture to enhance soil biological functioning and crop performance, yet their field effectiveness remains variable. This study evaluated how soil resource status, extracellular enzyme activities, and plant traits were associated with maize grain weight per plant under no-tillage straw mulching. A two-year field microplot experiment was conducted in an irrigated semi-arid maize system in northeastern China, comparing four microbial inoculants (M44, CF, SF, and YM) with an uninoculated control (CK). Maize growth and reproductive traits, soil physicochemical properties, and extracellular enzyme activities were measured, with soil variables assessed at depths of 0–10 and 10–20 cm. Microbial inoculation significantly affected aboveground biomass (p = 0.013), hundred-grain weight (p = 0.003), and grain weight per plant (p < 0.001), but not ear weight (p = 0.083). CF and M44 produced the strongest plant-level responses: their two-year mean aboveground biomass exceeded CK by 13.24% and 10.25%, respectively, while grain weight per plant increased by 10.78% and 10.66%. SF showed a weaker grain-weight response but increased β-glucosidase activity by 35.3% relative to CK in 2024, whereas YM increased available phosphorus by 70.8% in 2025 without a comparable grain-weight response. Soil treatment responses varied between years, with significant effects on soil organic carbon and available phosphorus in 2024 and on available phosphorus and soil water content in 2025. Treatment effects on β-glucosidase and cellobiohydrolase were detected only in 2024, while phosphatase was unaffected in either year. The random forest model explained 63.6% of the out-of-bag variation in grain weight per plant (p ≤ 0.001), with plant growth and reproductive traits showing the greatest predictive importance; soil organic carbon, soil water content, and cellobiohydrolase provided additional predictive information. Overall, inoculant responses were product-specific, and variation in grain weight per plant was jointly associated with plant development, soil resource status, and microbial functioning.