Discharge prediction is fundamental to hydrology, particularly in reducing the risks of flash floods in rural and urban areas. To address this challenge, the present study explores the potential of various Artificial Intelligence (AI) techniques for accurate daily discharge forecasting. Using historical daily discharge data from the Ravi River, which includes 2919 data points, the study involves comprehensive data preprocessing to eliminate outliers, ensuring the accuracy of the analysis. The preprocessed data is then analyzed using a range of machine learning and deep learning models, including Artificial Neural Networks (ANNs), Convolutional Neural Networks (CNNs), Long-short-term Memory (LSTM), and XGBoost (XGB). The performance of these models is assessed through multiple statistical metrics, such as R2, RAE, RMSE, RRSE, MSE, MAE, KGE, and VAF. The results indicate that the CNN model outperforms the other models, achieving a correlation coefficient of 0.821, highlighting its superior predictive capability. This study underscores the effectiveness of AI techniques in enhancing discharge prediction systems for flood risk management.