Southern Africa (SAF; Africa south of 10°S) summer monsoon (SAFSM) rainfall that usually occurs during December-Juanuary-February (DJF) season is the life-blood for agricultural and hydropower sectors, the backbone for most economies in SAF. This study was proposed to advance our understanding on SAFSM rainfall intraseasonal variability (ISV), and its prediction by the subseasonal to seasonal (S2S) prediction project models within DJF season. The study was split into two parts: diagnosing and modelling, whose results were based primarily on gauge daily data (1979–2005) and weekly models data (1999–2010), respectively. In first part, S-mode Empirical Orthogonal Functions (EOFs) with varimax rotation, applied to rainfall anomalies at different intraseasonal frequency bands (10–25, 10–90 and 20–90 days) revealed 4 most important and similar modes of variability in all analyzed bands. Although correlation analysis between timeseries of each rotated EOF (REOF) in all three bands and both low-level zonal and meridional wind’s components wind over SAF grid points showed the enhanced rainfall in all REOFs being associated with an anomalous cyclonic circulation,  it has been found that the anomalous circulation related with REOF3 and REOF4 appears to be more associated with variations in monsoon circulation (enhanced northwesterly, and northesterly monsoon flows, and even the southeast trade winds, besides the Mozambique Channel Trough), favourable for enhanced monsoon rainfall in the core monsoon region. Furthermore, the factor loadings of REOF3 and REOF4 shows together to encompass the monsoon core region over SAF (particularly Mozambique), although REOF3 retained for further analysis appeared representing alone most of the variability in the monsoon precipitation index (MPI) box. Power spectral density analysis applied to the total (10–90 days) intraseasonal REOF3 time series reaveled stronger statistically significant peaks at the 10–30-day band compered with those within the Madden-Julian Oscillation (MJO) time band (20–90 days) confirming thus the study hypothesis that the subseasonal summer monsoon rainfall anomalies over SAF, particularly, Mozambique are in part modulated by the MJO related convective activity. The pattern of lag composite anomalies keyed to positive and negative phases of the 20–90 days REOF3 confirmed it by showing a period of about 24 days for anomalies to complete a full cycle and anomalies resembling MJO phase 1 related convection some days before the rainfall hits Mozambique, suggesting that the existing teleconnection between South America and SAF is likely responsible for MJO impacts on Mozambique monsoon rainfall. To facilitate the prediction and monitoring of monsoon active and break days, a MPI and associated circulation indices were proposed. Through them the ability of all 11 S2S project models in simulating monsoon active and break periods over SAF for the commom S2S period (1999–2010) was assessed. Although the models can forecast the local large-scale zonal wind anomalies for lead times up to 3 weeks, predictive skill of monsoon rainfall anomalies appears limited to a week. The model’s rank showed ECMWF, JMA, UKMO, CNRM, KMA, and NCEP, as the top scoring ones, although they fail reproducing the MJO impacts on monsoon active periods. All retained models tend to reproduce the convective anomalies associated with observed monsoon active periods, although their associated circulation anomalies most of models with exception of KMA fail to reproduce. The latter is likely due to errors associated with models deficiencies in representing teleconnections.