In recent years, extreme weather such as drought and floods have occurred more frequently, and the study of watershed water cycle under changing environments is of great research significance. However, the construction of hydrological and meteorological stations in some regions of the world is not perfect enough, and the consistency of hydrological data in some regions is also disrupted due to changes in underlying surface and climate. These have led to lack of data in many study areas, posing a great obstacle to watershed hydrological simulation. As a multinational basin, the research data in the study area is difficult to fully obtain, which leads to similar difficulties in hydrological simulation of cross-border watersheds as in hydrological simulation of areas lacking data. A set of hydrological research methods suitable for regions lacking data in cross-border watersheds are aimed to be explored.The research method was mainly based on hydrological model and watershed similarity theory. Meteorological data, as an important input data for hydrological simulation, is often difficult to fully obtain in areas without data. Reanalysis data CMADS as a supplement to meteorological data in the hydrological simulation of the Red River basin was selected and its applicability in the Red River basin was studied. Based on the SWAT model and watershed elevation data, the Red River basin was divided into 123 sub watersheds. Then, a series of indicators such as elevation, slope, shape coefficient, forest coverage, precipitation and evapotranspiration were introduced to analyze the similarity of 123 sub watersheds by using of principal component analysis. According to existing little hydrological data and similarity results, four groups of similar watersheds were divided. Model parameters were well calibrated and validated in the basins with enough data, and then the validated parameters were used to drive the hydrological model in the data-shortage basins. The research results were as follows: In the Red River basin, the SWAT model was driven by both traditional meteorological station data and CMADS. Results showed that the CMADS-driven model performed well in the Son Tay station in both the calibration and validation periods, with the ERSR≤0.50 and ENS>0.75. Compared with the results of the CMADS-driven model, the model driven by traditional meteorological station data still satisfied simulation accuracy requirements, but the PBIAS was 12.1% in the calibration period and as high as 21% in the validation period, indicating a large simulation error. This indicated that CMADS could greatly enrich meteorological data in the Red River basin. Based on the results of basin similarity analysis and parameters transplantation, hydrological modeling in data-scarce sub-basins were carried out, which improved the simulation accuracy in most data-scarce sub-basins. Among them,ENS increased from 0.69 to 0.75 for sub-basin 77 after parameter transplantation from sub-basin 123,ENS increased from 0.75 to 0.85 for sub-basin 95 after parameter transplantation from sub-basin 123,ENS increased from 0.64 to 0.78 for sub-basin 96 after parameter transplantation from sub-basin 108, and ENS increased from 0.62 to 0.67 for sub-basin 102 after parameter transplantation from sub-basin 115. This indicated that parameter transplantation based on basin similarity could provide suitable parameters for hydrological models in data-scarce basins and provide more possibilities for hydrological simulation in data-scarce areas. The main conclusions were summarized as follows: Firstly, under the condition of lack of meteorological data, CMADS could be the alternative data for the SWAT model simulation. Compared with simulation results obtained by traditional meteorological station data, the results obtained by the CMADS-driven model could reach a very good standard in both the calibration and validation periods. Secondly, parameter transplantation based on basin similarity can provide suitable parameters for hydrological models in data-scarce areas, and basin similarity method improved the simulation results of most data-scarce sub-basins after parameter transplantation between similar basins. Finally, one research method was put forward successfully, which provide more possibilities for hydrological simulation in more data-scarce areas.