Heart disorders are a major health problem that affect the quality of life and can lead to serious consequences if not detected and treated promptly. In this context, the development of rapid and accurate diagnostic tools becomes extremely important. Electrocardiography (ECG) is a common diagnostic technique, but its analysis and diagnosis still face many challenges. To address this limitation, this paper proposes the use of convolutional neural networks (CNN) to classify heart disorders based on ECG signals. The research results show that the CNN model achieves high accuracy in detecting cardiovascular abnormalities, such as arrhythmia, myocardial infarction, and atrial fibrillation. This opens up the prospect of applying this method in clinical practice, helping to improve the diagnosis and monitoring of heart disorders. In the future, research may focus on model optimization, dataset expansion, and integration into automated diagnostic systems, to improve the efficiency and practical applicability of this method.