Cropland abandonment poses significant risks to both food security and ecological protection. Due to complex influences such as baseline land conditions, environmental factors, and socio-economic impacts, the process of land-use change associated with cropland abandonment is intricate. Especially, both the identification of longterm abandoned cropland and whether it has decoupled from socio-economic changes remain unclear. This study focused on the Loess Plateau in China to establish a set of rules for identifying abandoned cropland in different moisture zones. Then the spatial distribution of abandoned cropland over a 21-year period (2000-2020) was extracted. And the Tapio decoupling model was applied to analyze the relationships between economic development, population expansion, and cropland abandonment at both long-term and short-term scales. The results indicate a decreasing trend in abandoned cropland area, with a marked agglomeration effect, particularly in the semi-arid region. Gross Domestic Product (GDP) exhibited a more pronounced decoupling state with abandoned cropland than Population density (POP), and strong decoupling was the predominant state for POP over the study period. This study explored the relationship between cropland abandonment and climate variability, and could provide insights into sustainable land use and food security in China by offering foundational data and decision-making support for formulating differentiated cropland protection policies in ecologically vulnerable regions under socio-economic transformation.