Theoretical comparisons are made of different models for synthetic aperture radar (SAR) imaging of ocean waves. The velocity bunching model, distributed surface model, Lyzenga model, Canadian Center for Remote Sensing/Radarsat (CCRS/RADARSAT) model, and the Ocean Research and Engineering (ORE) model are compared analytically with respect to their description of the SAR imaging process and the radar backscatter process. The study concentrates on the temporal description of the backscatter process and its impact on SAR imaging of the ocean. Sample simulations are conducted so as to quantify some of the similarities and differences of the aforementioned models. Other SAR imaging models are discussed, as they pertain to the comparisons noted above. The analytical and numerical simulations show that the Lyzenga model, the CCRS/RADARSAT, model and the ORE model are similar for most commonly occurring sea states. The velocity bunching model is similar to these models when the SAR integration time is small, but it tends to deviate for long integration times. The distributed surface model is shown to describe the imaging process adequately when the radar backscatter is highly correlated in space and time. ¿American Geophysical Union 1990