<p>Floods in the southern Indian state of Kerala are the most frequent and deadliest types of natural disasters, leading to large economic losses and having killed over 410 people since June 2018. Despite Kerala's flood vulnerability and due to its intense annual rainfall, existing flood prediction approaches often fail to provide accurate and localized risk assessments. Various Machine Learning (ML) approaches offer promising results but there is a critical gap in applying these advanced ML algorithms along with systematized decision-making frameworks that would work well for Kerala's specific geographical and climatic conditions. This research aims to develop flood susceptibility maps for Kerala using a spatial flood database integrated within the ArcGIS interface. An efficient framework integrating Light Gradient Boosting Machine (LightGBM) learning models with the Analytic Hierarchy Process (AHP) is proposed to address this issue by enhancing flood susceptibility understanding and informed decision-making. The resultant map is validated using the Area Under Curve approach, which showed good accuracy of 0.948 and 0.763 for the LightGBM and AHP methods, respectively. The result indicates that the use of machine learning algorithm resulted better reliability of flood susceptibility maps. This study's findings apply to local authorities to maintain their preparedness and response efforts as well as minimize flood consequences.</p>