Originally developed and validated at the Cabauw (Netherlands) topographically flat onshore location, the alpha-I wind resource extrapolating method was tested at the FINO3 offshore site in the North Sea (Germany). The aim was to prove its validity also when applied over a substantially different environment in terms of surface characteristics and stability conditions. Data from local mast at 30, 80, and 100 m were used, with extrapolations to 80-m and 100-m turbine hub heights accomplished based on 30-m turbulence intensity observations. Trained over a 2-year period (2011-2012), the method was validated on year 2013. Similarly to the onshore application, the method was reliable in extrapolating wind speed to both 80 m and 100 m, with bias within 5%, NRMSE = 0.20 and r = 0.94. Conversely, scores were largely better than at the onshore site in predicting the annual energy yield, biased by 0.41-1.02% at 80 m, and 1.12-1.36% at 100 m. The method proved to be highly sensitive to the stability classification, as not considering this option increased its biases to 4.51-5.93% at 80 m, and 7.46-8.23% at 100 m. Method's reliability might suitably help reduce the number of masts installed throughout a large offshore area.