Abstract: Hydrological processes play an important role in the improvement of climate sys tem model. The spatial heterogeneous distribution of soil moisture content has a great influence on calculated evaporation and runoff as well as surface energy partition into latent and sensible heats. However，most of currently used land surface models (LSMs) are not able to take into consideration the influenc e of the heterogeneous distribution of soil moisture. The topographic index model (TOPMODEL) is a physicsbased model with a few parameters, less computer resourc es required and the capacity of taking into consideration the effect of topography on heterogeneous distribution of soil moisture by which the model is capable of dealing with the heterogeneous hydrological process of two dimensional characteristics in a land surface model. This index model has been recommended for use in current LSMs to improve the simulation skill of hydrological process on land surface. When the TOPMODEL is used for numerical studies, there could be different setting or selection for some parameters in several schemes therein such as spatial resolution of data base, Channel Initiation Threshold (CIT)，division of unsaturated areas in a catchment and spatial distribution of land surface slope, which will have important effect on the model simulated results. The way how to set the parameters in these schemes reasonably with no accuracy losing and reducing the requirement to the data and computer resources is very critical. In this study, a detailed numerical study of the schemes in the model has been conducted with the result that the effect of different setting for the schemes on the simulation outcomes of water balance in a catchment is revealed. The main conclusions from the study are as follows: (1)the simulated results of the case with the catchment being divided into 16 parts according to topographic index values are closer to those of the case with catchment being divided into only two (saturated and unsaturated parts) or three (saturated, wetter and drier parts) parts. So, when the TOPMPDEL is applied to LSMs, it may be a reasonable simplification to divide catchment into two (saturated and unsaturated parts) or three (saturated, wetter and drier parts) parts; (2) It is found from the study that the simulation outcome of water balancewill be distorted if the effect of slope for the catchment is ignored completely; however, the simulated result of water balance for the catchment with one meanslope given are closer to the result of water balance for the catchment with real spatial distributed slope set, suggesting that the way used in the current LSMs considering the grid as the one-dimensional homogeneous should be improved, and, on the other hand, considering topographic slope information in very detailed way may not be necessary. All the results provide the useful information for the development of a simplified and reasonably accurate hydrological model for use in a land surface model.