Abstract: The sustainable and stable development of cultivated land use system resilience holds paramount strategic significance for safeguarding national food security and facilitating agricultural green transformation. Conducting scientific analyses on the resilience of cultivated land use systems and the spatio-temporal heterogeneity of their constraining factors plays a pivotal role in mitigating ecological risks and stabilizing agricultural economic development. This study establishes a resilience evaluation index system for Hunan Province's cultivated land use system under the "Pressure-State-Response" (PSR) conceptual framework. Employing advanced methodologies including the Quantum Genetic Algorithm-Projection Pursuit (QGA-PP) model, kernel density estimation, and obstacle factor diagnosis modeling, based on the multi-scale "province-agricultural area-county", we systematically investigate the resilience evolution patterns and constraint mechanisms within Hunan's cultivated land use system. Results indicate:(1) Temporal Dynamics of Resilience: The resilience level of Hunan's cultivated land utilization system exhibited a phased fluctuation pattern across regions. The composite resilience index demonstrated an initial increase from 1.86 in 2008 to a peak of 2.18 in 2015, followed by a gradual decline to 2.06 by 2022; South-Central Hunan demonstrated marked resilience enhancement, while absolute gaps widened between South-Central and Western Hunan. The Changsha-Zhuzhou-Xiangtan metropolitan agricultural zone exhibited aggravated polarization, contrasting with convergence trends in Dongting lake basin agricultural zone. (2) Due to weak human disturbances and efficient disaster response governance, the Xiangxi hilly agricultural zone has become a highly resilient area. Meanwhile, the Central-Southern Hunan water-saving agricultural zone, benefiting from superior cultivated land resources and greater investment, has seen the high resilience values shift to this area. In contrast, the Changsha-Zhuzhou-Xiangtan metropolitan agricultural zone demonstrates weaker system resilience due to insufficient emphasis on arable land protection, resulting in limited capacity to resist disturbances and conflicts. (3) At the criteria level, the obstruction degrees across the province rank as follows: pressure layer > response layer > state layer. Notably, the obstruction degree of the response layer has shown a decreasing trend, while that of the state layer displays an increasing tendency. At the factor level, key obstruction factors include precipitation anomaly value, self-calibrated Palmer Drought Severity Index, per capita arable land area, agricultural fiscal expenditure per unit of cultivated land area, and agricultural disaster governance effectiveness. (4) Spatial heterogeneity of key constraint factors was observed across diverse agricultural zones. Based on this, a three-level differentiated resilience improvement strategy of province-agricultural area-county is proposed. The research proposes strategies for enhancing the resilience of differentiated cultivated land use systems, providing a scientific reference for the three-level synergistic cultivated land protection (province–agricultural zone–county) in Hunan Province.