In order to compensate for the missing report shortcomings of traditional forecasting methods for strong winds generated by mesoscale and microscale weather systems on the Nansihu Lake and minimize the loss of people and property, the Nansihu Lake Surface Strong Wind Early Warning System was designed. The intensive observing data every ten minutes from intensive automatic weather stations and conventional groundbased meteorological stations, the methods for extracting and analyzing realtime data, and drawing and releasing color contour maps, the sound alarming, etc., are fully used in the system to achieve the realtime observation of strong wind distribution and further improve forecast accuracy. The trial application shows that the system is interface friendly, easy to transplant, has high operating stability and low false forecast rate, which can play an important role in the shortrange early warning and forecasting of strong winds.