Abstract: In order to enhance the flood forecasting capability of small and medium-sized rivers in semi-arid areas with data scarcity, a comprehensive forecasting scheme formulation method is proposed to address the core challenges of strong flood suddenness and lack of hydrological data. This method is based on multi-source information fusion, and comprehensively utilizes historical data, terrain features extracted from digital elevation models(DEM), regional calculation, and parameter transplantation techniques to construct a solution for data deficient areas. Following the principles of "usability, maturity, and efficiency", a hierarchical and classified model adaptation system has been established, and specialized software tools have been applied to improve compilation efficiency. In the application practice of 30 sections of the Daqing River and Yongding River systems in Hebei Province, 92 sets of forecasting schemes were successfully constructed, verifying that the proposed method can effectively formulate accurate and qualified forecasting schemes. The results have important reference value for the formulation of flood forecasting schemes and flood control and disaster reduction practices in similar areas.