This study assesses the prediction performance of the East Asian Winter Monsoon (EAWM) using seasonal hindcast data(1993-2016) from the German Climate Forecast System(GCFS2). Main features of the EAWM are well predicted by the GCFS2,including the Siberian High, the East Asian trough , the East Asian jet stream, and the surface air temperature, and precipitation over East Asia.The interannual variations of East Asian trough and East Asian surface air temperature are skillfully predicted by GCFS2. GCFS2 shows prediction skills for the EAWM index (EAWMI) defined by sea level pressure. At the same time,the EAWM-related atmospheric circulation, surface air temperature, and precipitation anomalies over oceans are also well predicted .The high prediction skills of EAWM in GCFS2 are mainly due to the successful reproductions of the EAWM-ENSO relationship and the ENSO teleconnection. The correlation coefficient between EAWM and ENSO is -0.46 (1993 – 2016), which is stronger than that of observation. This means that the enhanced EAWM-ENSO relationship in GCFS2 is helpful to predict EAWM 2 months leading or longer. The EAWMI initialized in December GCFS2 still has 0.42 prediction skills after removing the ENSO signal, which indicates another source of prediction – the sea ice coverage in the Barents-Karabakh region in winter (BK_SIC) – works. The weaken of BK_SIC leads to the enhanced Siberian high pressure (SH) and the enhanced EAWM in the observation. The change of BK_SIC in the model can increase the predictability of northeastern SH, resulting in an increase in the EAWM prediction skills for December initialized.