The human microbiome operates as a complex, interconnected ecosystem where microbial interactions dictate community stability, host health, and disease progression. Understanding these dynamics requires moving beyond simple taxonomic catalogs to systems-level network analyses. This chapter reviews network methodologies in microbiome research, progressing from traditional correlation-based approaches to advanced artificial intelligence techniques. We systematically cover co-occurrence, protein-protein interaction, metabolic, multi-omics integrated, and evolutionary transmission networks. Computational tools—spanning general platforms and specialized pipelines—are compared alongside topology metrics and community detection algorithms. Furthermore, we highlight the integration of graph neural networks and protein language models, discussing current challenges in data standardization, model interpretability, and the merging of mechanistic and data-driven paradigms.