To promote the process of weather forecasting to be more professional and accurate, the potential forecast for predicting severe weather area has been conducted by Severe Weather Prediction Center in NMC since 2009. Realtime products for monitoring strong convective weather based on severe weather report from automatic weather stations, observations from national lightning monitoring network, FYseries satellite and radar networking etc. have been put in use. Criterion for mesoscale weather analysis has been established in MICAPS V3.0 system. The diagnostic products of mesoscale convective system outputted by the application of T639 global model， GRAPESRUC model and WRFEPS model have been developed. Guiding products for predicting classified severe weather (thunderstorm, hail and gale) area have also been issued. Prediction qualities were examined and assessed from April to September in 2009. The results show that the 6hourinterval TS scores of thunderstorm， local torrential rain， and hail and gale are 18%, 2.6%, and 2.1% separately, while 12hourinterval TS scores for each are 18.4%, 4.1% and 1.3%, respectively.