Abstract: By using the meteorological data observed by Mianyang Airport observatory hour by hour, sounding data and lightning location system data from Mianyang City, this article calculates the correlation coefficients between the thunderstorms and the convection parameters. The parameters with high correlation coefficients are selected as predictors and the physical meaning of each forecasting factor are discussed. By means of three methods of event probability regression analysis (REEP), Fisher discrimination and index superposition the potential forecast of thunderstorm in airport is studied. Among the three methods, the index superposition has the highest CSI score, and the thunderstorm occurrence probability can be determined by the magnitude of the accumulated value N. It tested the thunderstorms occur from June to August in Mianyang Airport by the index superposition, the CSI score is 52.9%, and the POD is 81.8%. It found that for thunderstorms around the airport on the route (municipal districts) has better forecasting ability by this method.