Artificial neural networks (ANNs) are computational models inspired by the structure and information-processing mechanisms of biological neural systems. This chapter introduces the basic concepts of artificial neurons, network architectures, activation functions, and learning paradigms including supervised, unsupervised, and reinforcement learning. The back-propagation (BP) network is presented in detail, covering its mathematical formulation, training procedures, convergence characteristics, and common improvements. Finally, a practical example of handwritten digit recognition demonstrates how BP networks perform feature mapping and classification. The chapter provides a foundational understanding of neural-network-based pattern recognition and its applications in machine vision.