The ac magnetostriction in electrical steel is commonly characterized in the time domain (e.g., the peak-to-peak, zero-to-peak amplitude) and also in the frequency domain (e.g., a harmonic analysis). However, due to the dynamical coupling of the test sample with the experimental set-up, the characterization of the magnetostriction (especially the one in the frequency domain) can give the wrong result. This research focuses on an experimental frequency characterization of magnetostriction and gives the theoretical background of the test sample’s dynamical coupling with the experimental set-up. The discussed natural dynamics of the test sample from the point of view of the different boundary conditions that can be used at the experiment gives a clear picture of the dynamical coupling. Besides the theoretical background, a detailed experimental approach is presented. This research theoretically and experimentally showed that the dynamical coupling of the sample can result in incorrect characterization of the magnetostriction. However, with the theoretical guidelines presented, the dynamical coupling can be completely avoided, which results in an accurate characterization of the magnetostriction.