The impact of flooding is exacerbated by unplanned settlements, particularly in developing countries. Addressing these issues involves mapping flood-prone areas and assessing their impact on populations and households. This study utilized datasets from Google Earth Engine (GEE), Food and Agriculture Organization (FAO), Central Bureau Statistics (CBS), and Earth Data to prepare maps of slope, drainage density, digital elevation model, rainfall, land use, and soil. These maps were created using Google Earth Engine (GEE) and QGIS through overlay analysis, considering factors such as influence and slope. Deep learning techniques were also integrated to enhance flood risk modeling, leveraging neural networks to analyze spatial and temporal data patterns. The risk assessment indicates that approximately twenty-four percent of the population is at high risk, with over three thousand settlements prone to flooding. The deep learning model revealed a significantly increasing trend of floods in the Morang district, providing more accurate and detailed risk predictions. The settlement risk map, enhanced by deep learning methods, can more effectively identify flood-safe and high-risk areas in the Morang district. This tool will aid local governments, urban planners, and communities planning residential areas to mitigate flooding risks.