The recent Spring Festival travel season has seen unusually complex weather conditions, but large-scale AI meteorological prediction models are revolutionizing the rules of weather forecasting. These models, through deep learning of historical data, accurately capture minute atmospheric changes, enhancing prediction accuracy under complex meteorological conditions. New-generation models such as FengWu-GHR and FuXi AI have improved the accuracy of global weather and ocean wave predictions. However, challenges remain, such as ensuring consistency with physical laws and addressing initial value perturbations.