This chapter provides a comprehensive overview of the foundational knowledge and frameworks in visual object tracking (VOT). It begins with an exploration of key deep learning models—Convolutional Neural Networks (CNNs), Multi-layer Perceptrons (MLPs), and Transformers—highlighting their structural designs, learning mechanisms, and roles in visual representation and temporal modeling. The chapter then introduces mainstream tracking paradigms including Siamese Networks, Discriminative Correlation Filters (DCFs), and One-Stream Frameworks, explaining their architectures, interaction strategies, and optimization objectives. These paradigms are discussed in the context of their adaptability to real-time and complex scenarios, such as occlusion, scale variation, and background interference. The evolution from handcrafted to learned features and from decoupled to unified modeling structures is emphasized. The chapter concludes by addressing current challenges and outlining future research directions. This foundational understanding supports readers in grasping advanced tracking algorithms and sets the stage for domain-specific discussions in later chapters.