<p>China faces the perennial challenge of ensuring food security due to its vast population and limited arable land. Against this backdrop, curbing uncontrolled non-grain use of cultivated land has become a critical policy concern. Although existing studies have examined the spatiotemporal patterns of non-grain conversion, the intrinsic temporal–spatial linkages and interactive driving mechanisms remain poorly understood. To address this gap, this study conducts a county-level analysis in Henan Province from 2012 to 2022, integrating spatiotemporal variance analysis, spatial autocorrelation modeling, and geographical detector techniques. The findings reveal that: (1) the non-grain conversion rate followed a U-shaped trend, declining initially before rising again, with marked regional disparities and pronounced clustering in the southwest and central regions; (2) temporal factors outweighed spatial heterogeneity in driving changes, with socioeconomic variables—particularly population density—emerging as the dominant influence; and (3) the interaction between population density and GDP exhibited the strongest explanatory power, underscoring the compounding effects of socioeconomic drivers. Based on these insights, the study proposes tailored policy measures to mitigate excessive non-grain expansion and safeguard sustainable grain production. It also contributes a novel spatiotemporal–mechanism analytical framework applicable to similar regional contexts.</p>