Water is among the most essential natural resources on our planet. However, effectively managing this vital resource presents a global challenge. The concept of Artificial Neural Networks (ANNs) draws inspiration from the intricate design of the biological nervous system, comprised of billions of interconnected neurons. This approach, rooted in machine learning, seeks to replicate the human brain's functionality. In the context of this study, three distinct models have been formulated, each utilizing lead times of 3, 6, and 9 h respectively. These models take input in the form of rainfall data from three specific stations, while the target data constitutes discharge information from a designated station. Following the construction of these models, an array of performance metrics, including RMSE (Root Mean Square Error), MAE (Mean Absolute Error), R-squared correlation coefficient, and NRMSE (Normalized Root Mean Square Error), were employed to assess their efficacy. Notably, the simulated discharge for the 9-h lead time aligns remarkably well with observed discharge levels, demonstrating strong agreement. Furthermore, the errors exhibited by the statistical indicators are notably minimal. The prowess of the ANN-driven model is evident, displaying proficient capabilities in accurate flood forecasting. This expertise makes it a valuable asset in the field of water resources engineering and management, providing a dependable tool to tackle these crucial tasks.