<p>Thin-walled beams are pivotal in modern lightweight and long-span structures due to their high load-bearing efficiency. However, their complex mechanical behaviors, such as distortion and warping, demand high-fidelity and efficient modeling methods. The published review literature has typically focused either on traditional deductive theories or emerging data-driven approaches in isolation. A unified critique of the interconnections and potential synergies between these two paradigms remains lacking. This review bridges this gap by providing a comprehensive analysis of their modeling methodologies. We systematically examine how cross-section deformation behaviors influence structural performance, and critically assess the applicability and limitations of each method. A particular focus is placed on how data-driven methods can recognize and explain traditional deformation modes. This work transcends a conventional literature summary; it offers a critical perspective on the future convergence of deductive and data-driven modeling methodologies. The findings provide a roadmap for developing next-generation, computationally efficient, and physically interpretable thin-walled beam models, serving as a valuable reference for the advanced design and engineering application of thin-walled structures.</p>