The experimental flow stress curves of structural steels obtained from axisymmetric compression tests conducted under hot-working conditions very often include the frictional effects present at the tool/specimen interface. Such effects have a significant influence on the flow stress and therefore, should be corrected prior to any quantitative analysis aimed at determining the constitutive description of these materials. Commonly, such a correction is carried out by assuming a constant friction coefficient (μ) or friction factor (m) independent of deformation conditions, which is an unrealistic approach. The present investigation analyzes experimentally the frictional effects that occur when steel is deformed under axisymmetric compression conditions in the temperature range of 850 to 1200 °C at a strain rate of 0.1 s−1 and applied effective strains of 1, employing cylindrical samples with an initial diameter to initial height ratio (d0/h0) in the range of 0.5 to 2. Finite element modeling (FEM), as well as element-free Galerkin modeling (EFGM), have been employed for the analysis and prediction of the von Mises stress distribution, barreling and amount of metal folding undergone by the compression specimens. It has been shown that the increase in flow stress due to frictional effects can be corrected on the basis of either μ or m, by assuming that these parameters vary in the course of plastic deformation and are strongly dependent on deformation temperature. A novel procedure for the systematic correction of the flow stress curves, taking into consideration the changes in friction conditions during plastic deformation, has been proposed.