Flood disasters are a serious threat to agricultural output, necessitating creative crop protection techniques. This study explores the use of machine learning algorithms, particularly the Random Forest model, in conjunction with rainwater-based civil infrastructure, like a pipeline system, to forecast and lessen flood disasters while protecting agricultural land. Rainwater is used to accurately anticipate floods by combining sophisticated modeling techniques with meteorological data, which greatly improves timeliness and dependability. Concurrently, well-placed civil constructions, such as pipes, serve as crucial infrastructures, controlling rainfall runoff and keeping important agricultural regions dry. Research results support the effectiveness of rainwater-based flood prediction. The Random Forest model performs exceptionally well, outperforming conventional methods with 94.8% accuracy and 92.5% precision. The ability of civil infrastructure to control water flow and protect crops from flooding is emphasized as a critical function they play in mitigating floods. In agricultural regions, a complete system for anticipating, avoiding, and mitigating flood disasters is formed by the synergistic integration of precipitation data and civil infrastructure. Case studies from various geographic contexts illustrate how flexible and effective these methods are, with a notable decrease in crop damage and an enhanced ability to withstand flooding events. This combination strategy’s benefits and drawbacks are assessed critically in the report, which also addresses issues and suggests directions for further study and use. This study emphasizes the effectiveness of Random Forest models with rainwater-based civil structures in flood forecasting and mitigation, offering valuable suggestions for integrating these approaches into agricultural and disaster management plans to safeguard critical resources.