Abstract: The intensity of the East Asian Summer Monsoon(EASM) has a close relationship with the distribution of rain bands and drought-flood patterns over China.To better do short-term climate prediction of EASM intensity, its multiscale variation characteristics and relationship with SST and 200 hPa zonal wind at interannual and interdecadal scales were studied using the wavelet transform method, Lanczos filter, and cross-validation.Subsequently, a multiscale statistical physical prediction model for EASM intensity, based on precursor signals, was constructed using the method of optimal subset regression.The results showed that EASM intensity exhibits quasi 4-, 13- and 43-year periodic oscillations. At the interannual scale, the SST in the eastern equatorial Pacific(10°N-10°S, 160°W-80°W) during the previous winter shows the largest significant negative correlation with EASM intensity, and has larger significant negative correlation with precursor signals in the 200 hPa zonal wind field.At interdecadal scales, the difference in 200 hPa zonal wind between approximately 60°S and 35°S has the largest significant positive correlation with EASM intensity.It also has larger significant positive correlation with precursor signals in the SST and 200 hPa zonal wind field, which includes the SST over the tropical Indian Ocean, low-latitude southeastern Pacific, and low-latitude southern Atlantic Ocean, and the 200 hPa zonal wind over the Asian subtropics.The potential of the two above-mentioned precursory factors in predicting EASM intensity was discussed, and the possible physical processes linking the EASM intensity and the two precursory factors at interannual and interdecadal scales explored.The multiscale optimal subset regression prediction model for EASM intensity was constructed with these precursor factors.The model not only showed better prediction ability for the interannual variation of EASM intensity, but also demonstrated certain predictive capability for extreme years.