This study aims to analyze the input data of the semi-distributed SWAT model that can be used as a hydrological study of the Lobo water reservoir. The adopted approach includes two major steps: The input data estimation and the watershed configuration. The estimation of input data consists in the collection and calculation of input data necessary for the watershed configuration. The input data collected and estimated were a digital elevation model (DEM), a soil map, soil physico-chemical parameters (grain size, K erodibility coefficient, exchangeable anions, soil density, organic carbon, soil electrical conductivity), a land cover map, information on agricultural practices and daily climate data (Temperature, Rainfall, Sunshine, Wind Speed and Relative Humidity). The watershed configuration consisted of the extraction of watershed boundaries and watershed network, the integration of data, the calculation of hydrological response units (HRU) and the watershed configuration. The SWAT model permitted to extract a single type of soil for Lobo water reservoir watershed. The watershed has a dense drainage system which means that the area is heavily drained. The degraded forest occupies 37% of the watershed followed by agricultural land (30% of the territory). 213 HRU from 39 sub-watersheds were also obtained. These data are an indispensable tool for the hydrological simulation of the Lobo reservoir.