Abstract: The performance of weather forecasting on the precipitation and typhoon movement in recent decades is analyzed. The results show that the forecasting skills have been obviously improved. In order to illustrate these kinds of achievements, the recent progress on operational numerical weather prediction models with the methodologies used in weather analysis and forecasting are discussed. The improvement of the forecasting skills mainly relies on the progress of the numerical predication model. The updated global spectral model with new cloud physics and high resolution is proved to be with good perform ance in anomaly correlation coefficient at 500 hPa over the Northern Hemisphere and higher scores in precipitation output. Satellite data assimilation is proved to be much efficient in model skills' improvement. Moreover, the synoptic analysis is still playing an important role in the routine weather forecasting based on the scientific understanding of weather related synoptic systems and the circulation patterns. The synoptic concept models on macro-scale weather systems and heavy rainfall related weather systems are derived from rainband movement analyses and torrential rain researches. The dynamic diagnosis and prognosis are coming to be important means in severe weather forecasting. An ingredient method is developed and used as a routine technique in heavy rainfall forecasting and convective weather forecasting. The statistic interpretation of model outputs can efficiently improve the routinely direct weather elements forecasting output. It is also a technical way to realize the high temporal and spatial resolution forecasting for temperatures and precipitations.