The web panels of metal sections subjected to compression have resistances that are often driven by local buckling. Transverse stiffeners are used to increase their resistance. These stiffeners can be arranged at different positions in relation to the edge of the metal section. The strength of the stiffened panel can be compared to that of a compressed element whose cross-section consists of the stiffener and a participating length of the web. To observe the behavior of these stiffened walls and to evaluate the existing analytical calculation methods, an experimental campaign was carried out on two metal sections of different heights, stiffened or not over their entire height of the web. Thus, 33 zones of stiffened or unstiffened panels are tested in double localized compression until failure. The stiffeners are placed near the edge or in the middle of the panel and welded on both sides of the web over its entire height and on the panel flanges. In parallel, a finite element model using shell elements and non-linear calculations considering the plasticity of the materials, the geometrical imperfections and the large displacements is set up. It is validated by comparing its force–displacement curves and modes of ruin with those obtained by tests. The model thus validated makes it possible to enrich the results of the experimental tests to better describe the behavior of stiffened walls subjected to compression. It also allows to discuss the observed failure modes and some analytical formulas which estimate the resistance capacities of the stiffened panels.