Research on the propagation of acoustic waves in ocean bottom sediment is of interest for active sonar applications such as target detection and remote sensing. Currently, all seabed scattering models available in the literature are based on the full solution of the wave equation, which sometimes leads to mathematically intractable problems. In the electromagnetics community, an alternative formulation that overcomes some of this complexity is radiative transfer theory, which has established itself as an important technique for remote sensing. In this work, radiative transfer (RT) theory is proposed for the first time as a tool for the study of seabed acoustic scattering. The focus of this work is the development of a complete model for the interaction of acoustic energy with water-saturated sediments. The general geometry considered in this study consists of multiple elastic layers containing random distributions of inhomogeneities. The accuracy of the proposed model is assessed by rigorous experimental work, with data collected from random media in which acoustic properties such as the concentration and size of scatterers, background material, and the presence of elastic boundaries are controlled parameters. First, the ultrasound RT model is implemented for layers of finite thickness. The range of applicability of the proposed model is then illustrated using scaled experiments conducted at the Northwest Electromagnetics and Acoustics Research Laboratory (NEAR-Lab). Next, the model is applied to field data collected in a region with gassy sediments and compared to the formulation originally used to explain these data. Finally, insight into the emerging area of study of the time-dependent RT formulation is presented, and its role in the representation of finite broadband pulses is discussed.