Effectiveness of high and low resolution GCM information was analyzed using probabilistic diagnostic method for Korean water resources managements. The formulation based on the significance probability of the Kolmogorov-Smirnov test for detecting differences between target (observation) and indicator variable (GCM). AMIP-II (Atmospheric Model Intercomparison Project-II) type GCM simulations done by ECMWF (European Centre for Medium-Range Weather Forecasts) were used for high resolution indicator variable and SMIP(Seasonal Prediction Model Intercomparison Project) type GCM simulations named Metri- AGCM(4°×5°) done by Korean Meteorological Agency (KMA) were used for low resolution indicator variable. The former has 2 and 2 degrees in longitude and latitude respectively and the latter has 4 and 5 degrees. Nodal surface precipitation and temperature values of both GCMs near 7 major river basins in Korea were used as indicator variables with analysis window concept. Observed mean areal precipitation and discharge values on each watershed were used for target variable. Monte Carlo simulations were used to establish the significant threshold of the estimator values. The results show that high resolution GCM is more significant to discriminate the extremes from target variables. It means that high resolution GCM can give more helpful information for water resources planning and managements. Considering this effectiveness, high resolution simulations are suggested for the future water resources management application in spite of various limitations of the present GCM simulations. Copyright ASCE 2005.