Historical Analysis of Land-use Changes in Vietnam’s Red River Delta: Bayesian Network Approach to Land Policies and Sustainable Development

[Purpose] Vietnam’s Red River Delta experiences rapid and uneven land-use transformations driven by market liberalization and urban expansion, severely impacting agricultural land and rural livelihoods. Accurate modeling of these changes is critical for sustainable land governance. To address this gap, this study uses Bayesian network modeling to retrospectively investigate LUCC in the Red River Delta, clarifying how land-policy transitions and agricultural expansion have influenced land-use decision-making and highlighting sustainable development implications. [Design/methodology/approach] This study proposes a Bayesian network-based panel decision support framework that synthesizes (i) multi-temporal satellite-derived spatial data (1979–2022), (ii) farmer land-use decision behavior, and (iii) historical land-policy and institutional change to evaluate and project LULC dynamics in the Red River Delta. To the best of our knowledge, this is the first region-wide, long-horizon application that explicitly links LUCC trajectories to policy shifts, quantifies transformation trends, and identifies the key drivers shaping land-use change. [Findings] The Red River Delta has undergone a clear shift from rice-based agriculture toward urban–industrial land uses between 2008 and 2022. Agricultural land declined sharply (7%), while forest land decreased only modestly (0.8%) and pastureland expanded (6.3%). The Bayesian network results indicate that industrial land prices are among the most influential economic drivers of these transitions, while the strongest governance levers relate to land-use zoning and conversion controls that steer agricultural-to-urban/industrial reallocations. In addition, the slight rebound of previously diminishing undefined agricultural zones suggests a move toward more structured land management. [Research limitations/implications] Limitations include medium-resolution satellite imagery potentially overlooking small-scale features and classification uncertainties from traditional algorithms. [Practical implications and Originality/value] This study presents a Bayesian network panel decision support framework that enables ex ante policy evaluation of land governance in the Red River Delta by simulating policy scenarios before implementation and estimating their likely effects on LUCC. It tests how changes in land use zoning, conversion controls, and industrial land prices shift the probability of major transitions, especially the conversion of rice-based agricultural land to urban and industrial uses, and highlights the most influential governance levers. The region-wide, long-horizon application in the Red River Delta offers a transferable approach for Asian delta regions facing similar trade-offs between urban industrial expansion, agricultural protection, and sustainability goals.