This paper describes the development of a streamflow forecasting model based on the the Sacramento Soil Moisture Accounting Model and applies optimal filtering techniques to sequentially update watershed-scale soil moisture state values, to improve streamflow predictions. In general hydrology is the study of the waters of the earth, especially with relation to the effects of precipitation and evaporation upon the occurrence and character of water in streams, lakes and water on or below the land surface. The Sacramento model is a hydrologic simulation model developed by the National Weaterer service, and used throughout the nation for operational streamflow forecasting. Here optimal filtering techniques are used in order to predict hydrological variables. An H-infinity filter is used to update daily estimates of the water content in model states representing watershed-scale soil moisture storage. Updated soil moisture storages are then used to predict daily streamflow. The output from the estimator is then compared with the model output without state updating.