In the era of digital transformation, network optimization has an important role in ensuring efficient, reliable, and scalable communication systems. The rise of 5G, the Internet of Things (IoT), and cloud computing increases complexity of networks and as a result traditional optimization methods struggle to meet the growing demand for speed, security, and resource efficiency. To mitigate this complexity, Machine Learning (ML) has emerged as a powerful tool to enhance network performance by enabling intelligent decision-making, predictive analytics, and real-time adaptation to dynamic conditions. To improve various aspects such as traffic management, resource allocation, fault detection, and security, machine Learning in network optimization leverages advanced algorithms. By analyzing vast amounts of network data, ML techniques can identify patterns, predict congestion, automate configurations, and optimize routing strategies with minimal human intervention. These capabilities lead to reduced latency, enhanced Quality of Service (QoS), and better utilization of network resources. This work explores the role of ML in network optimization, discussing key techniques, applications, and challenges. We examine how supervised, unsupervised, and reinforcement learning models are being applied to optimize network architectures and ensure seamless connectivity in modern digital infrastructures.