Abstract: Revealing the spatial correlation network of agricultural ecological resilience and its evolutionary mechanisms provides an important scientific basis for addressing climate change and enhancing regional capacity for coordinated adaptation. Based on the pressure-state-response framework and panel data from 31 provincial-level regions in China from 2014 to 2024, this study measures agricultural ecological resilience using the entropy weight method and constructs a binary spatial correlation network using a modified gravity model. Social network analysis, the temporal exponential random graph model, and the separable temporal exponential random graph model are employed to characterize network structure and examine the factors influencing the formation and persistence of network ties. The results show that: 1) The spatial pattern of the network changes in distinct stages, while the major overall structural indicators exhibit no significant monotonic trends. 2) Network density fluctuates over time and is higher at the end of the study period than at the beginning. Overall connectivity increases, while the network's hierarchical structure exhibits considerable interannual variation. The network core expands from several western provinces to multiple provinces in central and western China, although some peripheral provinces remain weakly connected. 3) Within-block ties predominate, while cross-block ties increase and block membership is reorganized. 4) Reciprocity and transitive closure are positively associated with both tie formation and persistence, whereas the shared partner effect is negative. The network also exhibits positive temporal dependence. 5) Under the baseline specification, climate adaptation capacity, extreme low temperatures, and geographical adjacency exhibit a dual-driving pattern, agricultural insurance exhibits a stabilizing pattern, and extreme heavy precipitation exhibits a tie formation inhibition pattern. Some results are sensitive to the binarization threshold. The classification of these functional roles is an exploratory statistical summary and does not constitute causal evidence.