In this work, we numerically simulated hydrogen thermal desorption spectroscopy on model martensitic steel samples using finite element analyzes. We studied the influence of trapping parameters (activation energy and untrapping frequency) and the effects of the sample size on the thermal desorption spectra. We discussed the final temperature of hydrogen desorption as a function of the sample thickness. We then determined the relevance of a low untrapping frequency and we discussed the choice of the most appropriate heating rate to correctly extract the hydrogen concentrations. We also questioned the critical size of the sample for the calculation of the activation energies using Gaussian deconvolutions for the common Choo-Lee plots. Copyright (C) 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.