Cracks often occur in beam-like structures, which are prone to shorten the structures′ service life and decrease their reliability. Accurately locating the crack damage is of great importance for the rapid and accurate maintenance of beam structures. In light of this problem, a crack location method is proposed for beam-like structures based on S transformation and mutual information entropy.First, the vibration signals are measured and transformed into time-frequency domain by S transformation.Then, the mutual information entropy of the time-frequency signals is computed before and after the beam-like structures are damaged.The damaged area is located by retrieving the minimum component in the mutual information entropy vector.In order to avoid the influences of the additional mass of measurement system, the machine vision system is employed to measure the distributed vibration responses of cantilever beams and composite blades. The damage locations are successfully identified by the proposed method and the effectiveness of this method is verified. This method can enrich vibration-based methods for structural damage detection in engineering applications and academic research.