A new approach for the validation of altimeter-derived sea state parameters using data from wind and wave models is proposed. Wind speeds and significant wave heights (SWH) acquired with the Geosat altimeter over the North Atlantic are compared with the wind analysis from the European Centre for Medium-Range Weather Forecasts (ECMWF) atmospheric model and with wave hindcast from the VAG model for the week of October 12--19, 1987. The statistical results with this 1-week data set agree with those of a 7-month comparison study between Geosat and the 43 National Data Buoy Center (NDBC) buoys. An advantage with the models is the large number of comparison points which allows an investigation of high sea state conditions and statistical analyses by binning in various ways. The underestimation of SWH by the Geosat altimeter, compared with the VAG hindcast, is confirmed and underestimation is shown to be proportional to the SWH. In order to use model hindcasts to validate altimeter-derived wind parameters, a method to improve the model wind data set is proposed. This method relies on the SWH comparisons between altimeter and model that are simultaneously available with the wind data. Tests on the 1-week data set show the validity of this new method. Applied to the three algorithms, Brown, smooth Brown, and Chelton-Wentz, these tests show that the statistical differences between the various algorithms are a function of wind speed range and indicate that the wind speeds derived with the Brown algorithm (which already agree best with buoy data) also agree best with the ECMWF 10-m wind fields. ¿ American Geophysical Union 1992