Abstract:[Objective] The patterns and mechanisms of farmland conversion to non-agricultural were explored in the hilly regions of Northeast Guangdong Province (NGP) to provide theoretical guidance for preventing Farmland Conversion to Non-agricultural. [Methods] Using NGP as an example, information on farmland conversion was extracted land-use images for the years 1990, 2000, 2010, and 2020. Spatial autocorrelation analysis and optimal parameter geographic detector methods were employed to investigate the spatiotemporal distribution characteristics and factors influencing farmland conversion. [Results] ① From 1990 to 2020, farmland conversion to non-agricultural in NGP exhibited a trend of “steady development followed by a sharp decline”, with a cumulative non-agriculturalized area spanning 2.29×105 hm2 at a rate of 25.31%. Forestland and constructed land were identified as the primary types of non-agricultural land. ② Vertically, croplands above 1 000 m in elevation and with a slope of 25°—35° were more prone to non-agriculturalization; horizontally, croplands within 50 m of construction land and 1 000—1 500 m from rivers were more likely to undergo non-agriculturalization. ③ Cropland non-agriculturalization in NGP exhibited significant positive clustering, with Global Molan’s I gradually decreasing from 0.371 to 0.255. The changes in “high-high” and “low-low” clustering dominated the evolution of the spatial autocorrelation pattern of cropland non-agriculturalization rates. ④ Agricultural population and cropland fragmentation consistently had strong explanatory power for the spatial distribution of cropland non-agriculturalization. In contrast, the explanatory power of socioeconomic factors for non-agriculturalization has weakened over the past decade. The interaction of multiple factors, especially agricultural and socioeconomic factors, can further enhance the explanatory power of cropland non-agriculturalization. [Conclusion] From 1990 to 2020, the area and spatial clustering of farmland conversion to non-agricultural in NGP decreased. However, the impact of the agricultural population and farmland fragmentation on farmland conversion to non-agricultural has intensified. It is suggested to reasonably control the speed of urban expansion, consolidate fragmented farmland, introduce small-scale agricultural machinery to improve farming conditions, and encourage farmers to cultivate by promoting rural e-commerce to address the issue of farmland conversion to non-agricultural.