The Electrocardiogram (ECG) is a non-invasive, cost-effective diagnostic technique for diagnosing cardiac diseases that monitors the heart’s electrical activity. ECG signal interpreting is challenging due to noise, distortions, and patient variability resulting in delays, and errors in diagnosing heart disorders. The objective of this work is to improve the accuracy and reliability of ECG-based Arrhythmia identification through the use of advanced AI methods. The ResNet-50 (DL model) and Decision Tree (ML model) are used to enhance signal accuracy and identify heart abnormalities such as Normal (N), Left Bundle Branch Block (LBBB), Right Bundle Branch Block (RBBB), Premature, Premature Ventricular Contraction (PVC), Atrial Premature Beats (APB). The models are trained on the MIT-BIH Arrhythmia dataset, which includes Temporal, Wavelet Transform and Heart Rate Variability features to capture heartbeat patterns. The ResNet-50 model performs better than the Decision Tree in terms of Accuracy (99.71 %), Precision (99.08%), Recall (98.96%), Specificity (99.82%), F1-Score (99.01%), AUC (0.99), AUPRC (0.97). The proposed models showed potential for improving ECG-based cardiac diagnostics by enabling the efficient and accurate automated identification of heart irregularities.