The hydrodynamic characterization of AUVs is essential for accurate dynamic modeling and control system design. This chapter explores experimental and data-driven approaches for determining hydrodynamic coefficients, which govern an AUV’s maneuvering behavior. Experimental methods, such as captive model tests (CMT) in towing tanks and free-running tests in open water, provide foundational data through standardized procedures like the Planar Motion Mechanism (PMM) and Rotating Arm (RA) setups. Recent advancements, including the Intelligent Towing Tank (ITT) at MIT, integrate artificial intelligence to enhance testing efficiency. Additionally, system identification techniques–both parametric (e.g., Kalman Filters, Support Vector Machines) and non-parametric (e.g., neural networks, Gaussian Processes)–offer data-driven alternatives that capture real-world dynamics more effectively. By reviewing these methodologies, this chapter highlights the evolution of hydrodynamic coefficient estimation, from traditional experimental techniques to modern machine learning-based approaches.