Abstract: Based on the adaptive network fuzzy inference system (ANFIS), the ways and methods to filter out noise of the impact factors from the main signal are discussed. Aiming at the abnormal weather events in 2010, with the delay-relevant method, three members of the summer monsoon system which most significantly affected the subtropical high anomalies in 2010 are analyzed from the observational data. They are the Mascarene cold high index, Somali low-level jet and the Indian monsoon latent heat flux. Because of adaptive learning and nonlinear superiority of the adaptive-network-based fuzzy inference system (ANFIS), it can be used to analyze and detect the influence and contribution of the members of the EASM (East Asia Summer Monsoon) system on the Western Pacific subtropical high (WPSH) anomalies. With the combination of the genetic algorithms and the statistical-dynamical reconstruction theory, a nonlinear statistical-dynamical model of the WPSH and three impact factors are objectively reconstructed from the actual data of 2010, and also a dynamically extended forecasting experiment is carried out. The results show that the forecasts of the subtropical high area index, the Mascarene cold high index, the Somali low-level jet and the Indian monsoon latent heat flux all have good performance in the short and medium ranges (less than 25 d). Not only is the forecasting trend accurate, but also the root mean square error is no more than 10%. Our paper not only provides new thinking for research on the association between the WPSH and EASM system, but also provides a new method for the prediction of the WPSH area index.