Accurate short-term wave forecasts are key for coastal activities. These forecasts rely on wind and currents as forcing, which in this work were both enhanced using neural networks (NNs) trained with satellite and radar data. Tested at three European sites, the NN-corrected winds were 35 % more accurate, and currents also improved. This led to improved IBI (Iberian–Biscay–Ireland) wave model predictions of wave height and period by 10 % and 17 %, respectively; even correcting under extreme events.