Abstract: Low-frequency rainfall over the lower reaches of the Yangtze River (valley) (LYRV) and the principal component of the circulation are adopted to establish a hybrid forecasting model with the multivariable lagged regressive model (MLR) and the principal component-complex autoregressive model (PC-CAR) combined, called MLR/PC-CAR model, which is applied to the daily forecasting of low frequency rainfall over LYRV for the extended range with the forecast period of validity prolonged. By many forecast experiments in June-August of 2011, this forecasting model has good predictive skill up to 50 days for the 20-30 day rainfalls over LYRV. And the predicted low frequency rainfalls over LYRV with the predictor of the principal component of 850 hPa meridional wind anomalies over the middle and high latitudes of the Southern Hemisphere is more accurate than that over East Asia, suggesting that, on the time scale of 20-30 days, the rainfall over LYRV are more tied to the principal components associated with the SCGT for the lag time. Moreover, the forecasting experiments for many years with the stronger 20-30 day oscillation also show that the SCGT is a key signal for the prediction of the low frequency rainfall in LYRV over the next 50 days with this hybrid forecasting model of MLR/PC-CAR. Based on the development and evolution of the SCGT, it will help us to hold the process of the anomaly change of a sharp turn from drought to flood in early June and the lasting heavy rainfall in mid-July of 2011 over LYRV. Hence, the variability of the SCGT is one of the main sources of the predictability for the extended range forecast of the 20-30 day rainfall and severe rainfall over LYRV in summer.