Abstract: Driven by economic incentives, agricultural restructuring, and regional resource endowments, the conversion of cropland to forest or fruit production has become increasingly prominent in China, emerging as a significant manifestation of non-grain land use. This process not only reshapes agricultural production patterns and rural land-use structures but also affects national food security, ecological sustainability, and regional agricultural development. Using multi-temporal land cover data from 1990 to 2024, this study integrated natural, socio-economic, and locational indicators to investigate the spatiotemporal evolution and driving mechanisms of cropland conversion to forest or fruit production across China. The GeoDetector and MaxEnt models were employed to analyze spatial differentiation characteristics, dominant influencing factors, and regional heterogeneity. Based on suitability evaluation results, optimization zoning and differentiated governance strategies were further proposed to support the governance of non-grain cropland use and the coordinated management of food production and ecological conservation. The results show that: (1) From 1990 to 2024, cropland conversion to forest or fruit production in China exhibited a stepwise increasing trend, with the total converted area reaching 14.10 million hectares and the conversion rate reaching 7.17%. The increase became particularly pronounced after 2000. Spatially, the phenomenon was mainly concentrated in southern China and the northeastern forest regions, especially in the southern coastal and southwestern areas. The expansion of forest and fruit production gradually occupied traditional grain-production spaces and altered cropland use patterns in many regions. (2) The spatial differentiation of cropland conversion to forest or fruit production was jointly driven by natural environmental, socio-economic, and locational factors, showing significant regional heterogeneity. At the national scale, slope was one of the most influential factors and showed a significant positive relationship with the conversion rate, indicating that sloping cropland was more likely to be transformed into forest or fruit production systems. Interactions among factors consistently demonstrated enhanced explanatory power, suggesting that the conversion process was shaped by the combined effects of multiple drivers. Regional differences in dominant mechanisms were evident. In Southwest and Northeast China, natural environmental factors such as topography, temperature, and precipitation played dominant roles. In South China, socio-economic and locational factors exerted stronger influences, while the middle and lower reaches of the Yangtze River reflected the combined influence of both natural and socio-economic conditions. (3) Grain-suitable regulation units covered 11.53 million hectares and were mainly distributed in Southwest China, the middle and lower reaches of the Yangtze River, and South China. These areas should be prioritized for comprehensive evaluation before implementing orderly restoration and recultivation while retaining ecologically important patches where necessary. Ecological retention units covered 2.52 million hectares and were mainly concentrated in Southwest China. In these regions, ecological protection and eco-friendly land-use patterns such as agroforestry should be promoted to balance ecological conservation and farmers’ income growth. Overall, this study reveals the long-term evolution patterns, regional differentiation characteristics, and driving mechanisms of cropland conversion to forest or fruit production in China, providing a scientific basis for differentiated governance of non-grain cropland use and the coordinated advancement of food security and ecological sustainability.