The sea surface temperature (SST), ocean wind speed (OWS), and sea surface salinity (SSS) are fundamental variables for understanding, monitoring, and predicting the state of the ocean and atmosphere. The analysis of these ocean parameters from passive microwave satellite measurements requires a Radiative Transfer Model (RTM). In this study, we compare three ocean RTMs from 1.4 to 89 GHz. A data set of satellite observations from Soil Moisture Active Passive (SMAP) and Advanced Microwave Scanning Radiometer‐2 (AMSR2) collocated with surface and atmospheric parameters from ECMWF ERA‐Interim and Mercator reanalysis has been developed. The selected ocean RTMs are as follows: LOCEAN a physical model with parameters adjusted to L band measurements, FASTEM (FAST microwave Emissivity Model) a fast parameterized model, and RSS (Remote Sensing Systems) an empirical model fitting satellite observations. Global systematic errors between simulations and observations tend to increase with frequency and are generally higher at horizontal than at vertical polarizations. Then, the analysis focuses on the accuracy of the RTMs as a function of the key ocean variables, SST, SSS, and OWS. Major discrepancies are found at frequencies above 1.4 GHz, for OWS higher than 7 m/s, with the LOCEAN and the FASTEM models, with differences strongly increasing with increasing OWS. Cold SSTs are identified as a source of disagreement between the simulations and the observations, regardless of the model. This is a critical issue, especially at 6 GHz, which is the key channel for the SST analysis from satellite. The present study is the first step toward the development of a new physically based community model.