Integrating Artificial Intelligence (AI) and signal processing techniques has become a promising option for building more efficient and reliable seizure detection systems. Electroencephalography (EEG) is a widely used tool for diagnosing and monitoring neurological conditions, and it is particularly effective in detecting episodes of epileptic seizures. This paper proposes a comprehensive approach to improve the reliability of EEG signals for enhanced seizure detection. In this approach, Discrete wavelet transform (DWT) is used for denoising the signal by thresholding the coefficients. Furthermore, entropy-based features are extracted and fed to the artificial neural network (ANN). The effectiveness of this proposed method is evaluated by experimenting with the CHB-MIT database. Mean square error (MSE) and correlation coefficient have been used to assess the effect of denoising. The accuracy, precision, sensitivity, specificity, F1-score and Area Under the Receiver Operating Characteristic curve (AUC-ROC) score obtained by the classifier for classifying EEG signals into seizure and non-seizure segments are 96.62%, 96.7%, 86.26%, 99.25%, 91.18% and 0.927 respectively. Seizure detection is a vital aspect of healthcare, particularly in monitoring and treating a neurological disorder like epilepsy (WHO, 2023), (Beghi, 2020). The timely and correct diagnosis of seizure occurrences is critical in improving patient outcomes, guiding treatment methods, and increasing the overall quality of life. This work highlights DWT’s potential as a useful method for denoising EEG signals for seizure detection along with the discriminative power of entropy based features. This approach will aid in creating robust reliable systems for the detection of seizures in clinical settings.