Abstract: Recessive farmland use transition has opened a new pathway for agricultural carbon emission reduction. While existing studies have primarily focused on the impact of farmland use transition on agricultural carbon emissions, few have explored how recessive farmland use transition affects agricultural carbon emission efficiency, indicating a need to broaden the scope and depth of research in this area. Based on panel data from 31 provinces spanning 2007 to 2023, this study employs the entropy method and a super-efficient SBM model that incorporates undesirable outputs to measure recessive farmland use transition and agricultural carbon emission efficiency. Using two-way fixed effects and threshold regression models, it examines the impact and mechanisms of recessive farmland use transition on agricultural carbon emission efficiency. The findings indicate that recessive farmland use transition significantly enhances agricultural carbon emission efficiency, a result that remains robust under multiple robustness tests. Heterogeneity analysis reveals that the effect varies significantly across regions with different geographical conditions, economic development levels, and agricultural foundations, with particularly pronounced impacts in western, economically less developed, and agriculturally weaker regions. Mechanism tests demonstrate that recessive farmland use transition improves agricultural carbon emission efficiency mainly through three channels: promoting technological upgrading, accelerating technological progress, and optimizing industrial structure. First, it facilitates the upgrading of agricultural production technologies, reducing input factors and improving utilization efficiency, thereby raising agricultural productivity. Second, it encourages the adoption of mechanized and intelligent equipment, accelerating the practical implementation of technological progress in agricultural production. Third, it generates industrial linkage effects, driving regional industrial structure toward green and low-carbon adjustment. Threshold effect tests confirm that cultivated land area per capita plays a key moderating role, exhibiting a single-threshold nonlinear characteristic of diminishing marginal effects: as cultivated land area per capita increases, the promoting effect of recessive farmland use transition on agricultural carbon emission efficiency gradually declines. This implies that in regions with smaller per capita cultivated land areas, promoting recessive transition yields higher returns, whereas once a certain scale is reached, its potential for efficiency gains relatively diminishes. Accordingly, this study proposes that recessive farmland use transition should be guided according to regional conditions, with optimized allocation of farmland resources and enhanced penetration and application of low-carbon agricultural technology innovations. Through the synergy of technology-driven approaches and structural optimization for green farmland transformation, along with appropriately scaled farmland management and regionally differentiated carbon reduction strategies, China can effectively accelerate progress toward its “dual carbon” goals.