The model used in this study is the Xinanjiang rainfall-runoff model that forms part of river modeling system for simulation of the rainfall-runoff process in sub-catchments. The meteorological input data to the modelare precipitation and potential evapotranspiration. In order to calibrate and verify the model, the generated runoff of the Xinanjiang model is routed,with Muskingum method for surface runoff, and with the linear reservoir method for groundwater runoff. In this study, the first step is the presentation of a brief introduction of the original structure of the runoff generating component of the Xinanjiang model. Based on the soil moisture storage capacity distribution of the Baishuikeng river basin, the corresponding modified computational formulae of runoff generation are then presented. The use of daily data of multiple years and hourly event data were collected from published and unpublished sources of 1980 to 1988 for the Baishuikeng river basin. Monthly mean evapotranspiration data were collected from some reports of the Baishuikeng river basin. These data were transformed into daily data. Daily hydrographs for calculated and observed discharges were derived using the rainfall and observed discharge data and evaporation data as input to the computer program. Manual calibration is used in this study to obtain the optimum parameter values of the Xinanjiang model. The Xinanjiang model consists of fifteen parameters that have to be calibrated before application. In the daily modeling, the Determination coefficient (CD), and the error values are within acceptable range which indicate a good fit between the simulated and observed valued and a good water balance in the basin. It was observed that the calibration locates 75％ success rate of optimum parameter values. Further more, results for the selected flood events modeling indicate that, in most of the events for both the calibrated and verified results, the error values are either less than or equal to zero, indicating the time of occurrence the observed peak flow and the simulated peak flow coincides. In the sensitivity analysis conducted, the statistical analysis supported the earlier work of Zhao; that the Xinanjiang model output values are more sensitive to these seven parameters (K, SM, KG, CG, CS and L). However, it is believed that the sensitivity analysis can be improved with better quality data. In this study, K and SM are the most sensitive parameters Based on the above, the overall result of this research is quite successful based on the fact that the model is a simple model with many parameters and limited available information. Finally, the results of this study could be used for the prediction of flood events and will help to manage reservoir operation in terms of flow prediction. It will also contribute to academic documentation for respective researchers.