The sustainability of rice fields is fundamental to maintaining regional food security, particularly in agricultural production centers such as Polewali Mandar Regency, West Sulawesi, Indonesia. Rapid population growth and spatial development have intensified pressure on productive agricultural land, increasing the risk of long-term decline in rice field availability. This study develops a spatial statistical modeling framework to predict the distribution of rice field conversion for the period 2026–2030 using an integration of Frequency Ratio (FR) and Spatial Multi-Criteria Analysis (SMCA) within a Geographic Information System (GIS) environment. Land cover data from 2010 and 2020 were used to identify historical conversion patterns and to construct predictive variables. Ten driving factors—including topography, slope, geomorphology, soil type, rainfall, accessibility, settlement characteristics, and spatial policy direction—were evaluated alongside protected agricultural zones as limiting constraints. Model validation using the Receiver Operating Characteristic (ROC) curve produced an AUC of 0.83 (success rate) and 0.75 (predictive rate), indicating good and reliable model performance. The projection results demonstrate continued pressure on rice field areas, particularly in zones influenced by infrastructure and settlement expansion. By providing spatially explicit predictions, this study offers a decision-support tool for proactive land-use regulation, agricultural protection policies, and strategic planning interventions aimed at safeguarding food self-sufficiency in the medium term.