The search for new parameters for predicting the all India summer monsoon rainfall (AISMR) has been an important aspect of long range prediction of AISMR. In recent years NCEP/NCAR reanalysis has improved the geographical coverage and availability of the data and this can be easily updated. In this study using NCEP/NCAR reanalysis data on temperature, zonal and meridional wind at different pressure levels, few predictors are identified and a prediction scheme is developed for predicting AISMR. The regression coeffcients are computed by stepwise multiple regression procedure. The final equation explained 87% of the variance with multiple correlation coeffcient (MCC), 0.934. The estimated rainfall in the El-Nino year of 1997 was -1.7% as against actual of 4.4%. The estimated rainfall deficiency in both the recent deficient years of 2002 and 2004 were -19.5% and -8.5% as against observed -20.4% and -11.5% respectively.