This study was conducted in Barru Regency, Indonesia, a region characterized by diverse topography, high agricultural potential, and environmental constraints. The aim was to assess the predictive performance of Random Forest (RF), a machine learning algorithm, for FAO-based land suitability classification and rice productivity estimation by integrating Geographic Information Systems (GIS) and a novel managerial variable—technical irrigation. Using 12 GIS-derived soil parameters, the RF model achieved a high accuracy of 0.95 ± 0.01 for land suitability classification through cross-validation. However, the model showed a lower performance in predicting rice productivity (0.32 ± 0.04), likely due to the complexity and noise of the agricultural data. The inclusion of irrigation data as an additional input variable increased productivity prediction accuracy by 6%, from 0.32 to 0.38 ± 0.06, highlighting the importance of managerial factors in enhancing model reliability and practical utility in tropical, infrastructure-limited areas. Permutation-based sensitivity analysis revealed that slope, cation exchange capacity (CEC), and soil depth were the most influential factors for land suitability, while slope, potassium, and CEC were key to productivity prediction. The resulting land suitability map indicated that class S3 (marginally suitable) dominated the study area (34,836.62 ha), followed by not suitable (N: 8,303.68 ha) and moderately suitable (S2: 2,722.62 ha), with Barru Sub-district having the largest S2 area (1,181.16 ha). These findings suggest that enhancing irrigation infrastructure and improving soil conditions in targeted areas can support better land management strategies. Although the model showed limitations in predicting productivity due to data complexity, the integration of managerial variables represents a novel approach that improves model performance and provides practical insights. This study contributes to the growing field of GIS and machine learning applications in precision agriculture and offers a scalable framework for sustainable land-use planning in tropical regions.