A framework based on GMDH (Group Method of Data Handling) is proposed in this paper to establish the I/O model as the alternative by using the relatively simple field measuring water/tidal level and rainfall data as the model input to predict 1 to 6 hour-lead water level of specific river during a storm period. The update stream level and rainfall data were collected to organize the Sequential GMDH modified model to match the time variant properties in stream level forecasting steps. The distance ~ level relation model based on the Muskingum formula and range data which were simply measured from a map were established in this paper also. The GMDH stream level forecasting model cooperated with the distance ~ level relation model could be used as the coupled model for 1 to 6 hour-lead water level forecasting at any specific stream sections.