The Chennai region's meteorological conditions are causing floods in many districts to occur more frequently and with greater intensity. Therefore, anticipating and planning for floods under severe weather conditions is essential for making decisions and handling impending calamities. These days, deep learning (DL) methods are crucial for supporting meteorological applications and successfully preventing natural hazards. The use of climatological data to enhance flood prediction is covered in this manuscript. By examining the effects of rainfall changes in several Chennai regions, the current study attempts to provide an accurate estimate of flood risks. Pre-processing and flood forecasting are the two stages that are completed for the proposed framework. The MaxAbsScaler (MAS) approach, which could be used to eliminate unwanted missing values from the database, is described in the preparation step. In order to predict the flood in various Chennai regions, the Extended Elman Spiking Neural Network (ExESNN) technique is then suggested. In order to avoid network problems during the training phase, the derived model's parameters are tuned using the Chaotic Artificial Hummingbird Optimizer (Ch-AHO). The Python platform is utilized to process the built framework, and the experimentation process makes use of the real-time flood prediction database gathered between 2000 and 2023. Numerous computational metrics are assessed and differentiated from various research, including R2, mean absolute error (MAE), Kling-Gupta Efficiency (KGE), Nash-Sutcliffe Efficiency (NSE), root mean square error (RMSE), and calculation time. In comparison to several traditional research on rainfall forecasting in various parts of Chennai, the developed technique yields an overall R2 of 0.994, RMSE of 0.851, KGE of 0.968, NSE of 0.991, MAE of 0.50, and overall CT of 63.33s.