The potential of utilizing eigenfrequencies corresponding to edgewise bending modes for predicting the bending strength of timber is investigated. The research includes measurements of axial and transversal resonance frequencies, laboratory assessment of density, static bending stiffness and bending strength of 105 boards of Norway spruce of dimensions 45×145×3600 mm. It is shown that Eb,1, (MOE based on the eigenfrequency of the first bending mode) gives a higher coefficient of determination to the bending strength than what Ea,1 (MOE based on the first axial eigenfrequency) does. It is also shown that eigenfrequencies corresponding to higher bending modes can be used in the definition of a new prediction variable, the modulus of inhomogeneity (MOI). This is a scalar value representing the lack of fit between the true, measured eigenfrequencies and the expected (assuming homogeneity) eigenfrequencies of a board. The results show that using the MOI as a third prediction variable, in addition to Eb,1 and density, increases the coefficient of determination with respect to bending strength from R 2 = 0.69 to R 2 = 0.75. Machine strength grading of timber based on dynamic excitation of boards has won large market shares in the last decade. The vibration content is detected using a microphone or a laser vibrometer and fast Fourier transformation is used for calculation of eigenfrequencies (also called resonance frequencies or natural frequencies) corresponding to axial modes of vibration. The common way of utilizing the information from the measured vibration content is to calculate the modulus of elasticity (MOE), or actually a mean axial stiffness, using the eigenfrequency corresponding to the first axial mode. One aim of this paper is to investigate how the dynamic board stiffness or MOE based on the eigenfrequency of the first edgewise bending mode correlates with the bending strength and compare with the correlation between the dynamic axial MOE and the bending strength.