Fractal-dimension-based signal processing has been extensively applied to various fields for nondestructive testing. The dynamic response signal can be utilized as an analytical tool to evaluate the structural health state without baseline data. The fractal features of the dynamic responses with fractal dimensions (FDs) were investigated using the Higuchi, Katz, and Sevcik methods. The waveform FD proposed by these methods was extracted from the measured frequency response function (FRF) data in the frequency domain. Damage was observed within this region, which resulted in an abrupt change in the curvature of the FD. The effectiveness of the methods was investigated via the results of a steel beam test and a numerical experiment to detect damage.