Cardiovascular diseases continue to be the primary contributor to deaths worldwide, highlighting the urgent need for enhanced approaches in diagnosis, prognosis, and patient monitoring. The adoption of deep learning has revolutionized cardiovascular healthcare by leveraging hierarchical feature extraction from raw clinical data, thereby enhancing outcomes in arrhythmia detection, cardiac image segmentation, and risk assessment tasks. This chapter reviews foundational deep learning architectures such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), autoencoders, generative adversarial networks (GANs), and transformers, which play distinct roles in analysing cardiovascular datasets. It also discusses public datasets such as MIT-BIH, PTB-XL, ACDC, and UK Biobank that have facilitated advances in model development and benchmarking. Applications across ECG classification, cardiac image analysis, EHR-based risk prediction, and edge AI-based remote monitoring are summarized. Furthermore, current challenges related to data quality, generalization, explainability, and regulatory compliance are examined. Future directions emphasize self-supervised learning, federated learning, multimodal integration, and digital twin models for personalized cardiovascular care. This chapter aims to support researchers and clinicians in harnessing DL to improve cardiovascular health outcomes.