Abstract: A short-term prediction ahead of time on monthly mean SSTA over Nino(1-4) oceanic region is made in the context of adaptive filtering using the Singular Spectrum Analysis(SSA) and Auto-regression(AR) model in this paper.The results show that the prediction ahead of time is very efficient to this 1997-1998 strong ENSO event and the simulational predictions ahead of time have high confidence to the three strong ENSO events in history record.Therefore,the forecast skills of this SSA-AR prediction scheme are steady and its independent sample test and real prediction possess high precision.If we revise further the SSA-AR scheme then it is possible that it becomes an efficient model of operational forecast for ENSO.