ABSTRACTHow boreal summer intraseasonal oscillation (BSISO) modulates the probability and spatial distributions of extreme rainfall occurrence over populous southern China was examined, using the newly proposedBSISOindices and two high‐resolution rain‐gauge‐based rainfall datasets in China. The probability density function of May–August rainfall in southern China is skewed towards large values in phases 2–4 of the first component and in phases 5–7 of the second component ofBSISOlife cycle, during which the probability of extreme rainfall events at the 75th (90th) percentile increases by 30–50% (over 60%) relative to the non‐BSISOperiod. The devastating floods with prolonged extreme rainfall in southern China over the three past decades occurred coincidently with theseBSISOphases. The first component ofBSISO, associated with 30–60‐day eastward/northeastward‐propagatingISO, is more favourable for the rainfall extreme over in‐land China. In contrast, the second component ofBSISO, related to the 10–30‐day northwestward propagatingISO, tends to link with the rainfall extreme along the southeast coast of South China. Moisture budget indicates that the favourable environment for rainfall extreme is associated with southwesterly moisture convergence over southern China, while the moisture advection contributes insignificantly. This study suggests a potential for monitoring and probabilistic prediction of extreme rainfall events in southern China based on the real‐timeBSISOindices.