Abstract: A grid-based distributed rainfall-runoff simulation system has been developed in this paper. The system can split the catchment into a number of square or rectangular grid elements using DEM, extract topographic attributes of elements such as slope, upslope flow contributing area, and classify the grid elements into hillslope elements and channel ones. Each hillslope grid element has model components for interception, infiltration, overland flow and lateral soil water flow, and each channel element has model component of channel flow. The infiltration and excess rainfall on each hillslope element are calculated using the Green-Ampt infiltration equation, soil lateral flows are based on Darcy's law and continuity equation, overland and channel flows are described by one dimensional kinematic wave approximation to the St Venant equations. Implicit finite differential scheme is used to solve these equations. The hydrological parameters are determined from digital soil and land use data. The system also has the ability of spatial data interpolation, visualization and statistics of the results. The interpolation approaches include Thiessen, Kriging, inverse distance weighted averaging, and trend surface methods. The visual components of system include the 1-D and 3-D visualization of rainfall, infiltration, surface runoff, and subsurface flows for one or all elements, and synchronized 3-D visualization for any two processes of the catchment.The system was applied to the simulation of two flood events happened in Hangtuling watershed, Zhejiang Province, China. The data input for the simulation were: rainfall and runoff time series of two events, the DEM, land use map, and soil map of the watershed. By sensitivity analysis of parameters, Manning's roughness of surface and soil saturated hydraulic conductivity were determined to be calibrated. The first event was used to calibrate the model, and the second for testing. The results show that the system is capable of simulating rainfall-runoff processes with least data requirements and easy operation.