Rainfall-runoff simulation hydrological models are generally applied to obtain synthetic series from historical rainfall records. A recent increase has been noticed in the demand for applying hydrological simulation models in watersheds in which there is an incipient gauge network or even no streamflow gauging at all. However, in the absence of streamflow data, rainfall-runoff models cannot have their parameters calibrated. Thus, it is necessary to use methods to transfer spatial information, generally referred to as hydrological regionalization, which is the approach used in this paper. The methods described here were applied to the Paraopeba river basin, a tributary of the Sao Francisco river, located in the Brazilian state of Minas Gerais. The hydrologic model used in this study is SMAP (Soil Moisture Accounting Procedure) which was calibrated using as objective function the minimization of the errors between the observed and calculated flows. For the regionalization method the multiple linear regression technique was used, relating the model parameters to the geomorphic, climatic, hydrogeological and soil characteristics of the basin. The adjustment equations, as obtained at other gauging stations, were performed using the Jack-knife method and from the performance indices found it can be concluded that the method can be used to evaluate, at least approximately, the availability of water resources in ungauged basins