This paper investigates the forward modeling of chirp-sonar data for the quantitative characterization of marine subbottom sediment between 1 and 10 kHz. The forward modeling, based on a transfer function approach, included impacts of layering or impedance mismatch, attenuation, roughness, and transitional layers, i.e., continuous impedance variations. The presented approach provided the best compromise between the number of available geoacoustic parameters from chirp-sonar data and the subbottom modeling accuracy. The forward model was tested on deep-sea chirp-sonar data acquired at a central frequency of 3.5 kHz. Comparisons between synthetic and experimental seismograms showed good agreement for the first 15 m of buried layers. Performance of the inversion using this forward model was also examined through sensitivity analysis. The results suggested that estimations of layer thickness, impedance, and transitional layer thickness were robust, whereas roughness and attenuation estimations were subject to wavelength and layer thickness conditions.