Extreme weather events have an important impact on economic and social development. The extreme weather events often result from superimposed influence of multi-factor and multi-scale abnormal signals of climate environment. Thus, extreme weather events embody the unity of weather and climate. This paper analyzes the difficulties and challenges of extreme weather event prediction from the perspective of the internal relationship between weather and climate, and also discusses the opportunities of extreme weather forecasting which are gained from the weather and climate integration models, multi-loop coupling assimilation and ensemble technology as well as the assimilation probability prediction technology based on deep learning and AI. Finally, the significance of the integration of weather and climate disciplines and the establishment of weather and climate integrated forecasting operation system is pointed out.