By means of Kalman filter, this study examines recursive updating of a distributed model's state variables, which were the water depths on a fine grid cell, using observed discharges at the basin outlet. The measurement update algorithm of the filter decides the optimal discharge, and the ratio of the optimal discharge to the simulated discharges was applied to every discharge from each cell for a reformation of the distributed state variables with a specific stage-discharge relationship. For the time update algorithm of the filter, Monte Carlo simulation was adopted for the prediction algorithm of the filter. The proposed algorithm offers efficient state variables updating in a distributed hydrologic model.