在原有相应涨差河道洪水预报模型的基础上 ,利用差分方法推导出仅用下游站流量或水位资料表达的改进型相应涨差模型。并用具有可变遗忘因子的递推最小二乘实时校正算法对模型参数进行在线识别。该模型用于长江枝城至监利河段的实际预报检验 ,证明模型结构合理 ,算法简捷 ,成果令人满意。 On the basis of the original flood forecasting model of the corresponding upwelling river, an improved corresponding upside difference model which can express the flow or water level data of the downstream station is deduced by using the difference method. The recursive least squares real-time correction algorithm with variable forgetting factor was used to identify the model parameters online. The model is applied to the actual forecast test from Zhicheng to Jianli reach of Changjiang River, which proves that the model is reasonable in structure and the algorithm is simple and the result is satisfying.