Abstract: Clarifying the spatio-temporal characteristics and driving factors of agricultural eco-efficiency (integrating carbon reduction, pollution abatement, green expansion, and growth) in Northeast China’s Three Provinces is crucial for advancing agricultural and rural modernization in the region. Utilizing panel data from 32 major cities from 2014 to 2023, this study employs the Undesirable Output Super-SBM model, kernel density estimation, spatial autocorrelation analysis, and spatial panel econometric models for empirical analysis. The findings indicate that: (1) The overall agricultural eco-efficiency in the region demonstrates an upward trend, yet inter-regional disparities exhibit phased characteristics, a widening tendency, and signs of polarization. (2) A significant positive global spatial autocorrelation is observed for agricultural eco-efficiency. Locally, spatial agglomeration manifests as five distinct types: cold spots, sub-cold spots, transition zones, sub-hot spots, and hot spots. (3) Improvements in urban agricultural eco-efficiency exhibit a significant negative spatial spillover effect. Specifically, a 1% increase in efficiency within a local city is associated with a 0.565% decrease in efficiency in neighboring cities. (4) Decomposition of spatial effects further reveals differences in the direct impacts and spatial spillover effects of various influencing factors. The study concludes that while a certain degree of spatial correlation exists, the strength and connectivity of these spatial linkages require enhancement. Recommendations are proposed, including strengthening inter-provincial and inter-city collaboration, implementing differentiated development strategies, and fostering in-depth cooperation across multiple dimensions to promote agricultural eco-efficiency.