In this paper CMOD5, a new C‐band geophysical model function (GMF), is derived on the basis of measurements from the scatterometer on board of the European Remote Sensing Satellite ERS‐2. First‐guess winds from the European Centre for Medium‐Range Weather Forecasts were used as a reference for the period from August to December 1998, adding up to more than 22,000,000 collocations. CMOD5 corrects some deficiencies of the currently widely used CMOD4 GMF. Linear and higher‐order wind speed corrections as computed with a triple collocation method are implemented. Recent measurements of extreme backscatter and wind obtained by aircraft and in situ data are fitted. Also, a more accurate fit of the two‐dimensional cone surface in three‐dimensional measurement space is established, especially in the regime of strong winds. These improvements result not only in better wind retrievals at high wind speed, but also in a more uniform performance across the ERS scatterometer swath. Moreover, the wind ambiguity problem has been reduced owing to the improved fit of the cone surface, resulting in about 75% skill of the first rank solution for winds above 10 m/s. These improvements aid the general usefulness of retrieved C‐band scatterometer winds for climate and weather applications, and the ambiguity removal in dynamical and extreme weather conditions in particular.