Based on the daily precipitation datasets at 200 stations in the south of China and the corresponding NCEP/NCAR daily reanalysis data from 1981 to 2008, Analysis By Synthesis (ABS) and Butterworth band-pass filter are adopted to diagnose the Low-Frequency Oscillation (LFO) characteristics of the summer rainfall over the Yangtze-Huaihe River Valley (YHRV). And then an experiment for the Extended-Range Forecast (ERF) of persistent heavy rainfall over YHRV is made. The major conclusions are as follows: summer precipitation over YHRV has an obvious characteristic of LFO with 20-50 days period, and for climate average, the 20-50-day low-frequency components account for nearly 20%, and there are significant annual variations for the intensity of LFO signal. There is a good relationship between the 20-50-day components and actual precipitation, especially the phase changes in peak values correspond to alternation of precipitation concentrated period and break period. Therefore the 20-50-day low-frequency components of precipitation which is taken as the predictand for the ERF experiment and the persistent heavy rainfall process over YHRV are closely linked. Meanwhile based on the links between LFO of large scale circulation in East Asia and summer persistent heavy rainfall over YHRV, 10 indexes of circulation critical systems are built, which can correctly reflect the intensity and tendency of the 20-50-day low-frequency components of precipitation, and are used as predictors for the ERF experiment. Combining with the NCEP Climate Forecast System (NCEP/CFS) output the forecast model based on the LFO signal has some reference value for ERF of persistent heavy rainfall over YHRV.