Amid rapid urbanization, the Beijing–Tianjin–Hebei region has experienced substantial land use restructuring. This spatial evolution has created conflicts between development and conservation, challenging regional sustainability. Using 2005–2020 data, this research incorporates urban land use efficiency (ULUE) and the closeness centrality of spatial correlation network into the Patch-generating Land Use Simulation (PLUS) model to reveal driving mechanisms. It simulates 2035 land use patterns under four scenarios: Business as Usual (BAU), Food Security (FS), Ecological Security (ES), and Efficiency-Driven Optimization (EDO). The findings reveal: (1) Regional land use transitioned from rapid expansion to structural adjustment, with cultivated land acting as the primary source for new construction land. (2) ULUE and the closeness centrality of its spatial correlation network exhibited high feature importance in land use transitions, confirming that land allocation efficiency and network structures significantly reshape regional patterns. (3) Compared to BAU or single baseline constraints, the EDO scenario effectively limits inefficient county expansion and promotes compact core city development, balancing construction land control with cultivated and blue-green space preservation. For large urban agglomerations, a spatial governance pathway integrating efficiency-driven strategies and stock optimization is optimal for coordinating multi-dimensional spatial objectives.