In this work a comparative study was carried out, in which different methods were used in the literature that seek to evaluate the number of stations and the quality of the information generated by the hydrometric network of a watershed, using Information Theory concepts. The underlying idea is the so-called optimal network whose function, according to World Meteorological Organization (WMO) is to optimally and inexpensively meet the primary goal of hydrometry, which is to provide the necessary information with a minimum number of stations correctly positioned in the basin. Methodologies based on Information Theory ascend to fill the gap on a standard method for the design of hydrometric networks. The evaluated methods were applied to the subbasin of the Rio das Velhas belonging to the São Francisco River basin in Brazil. The results showed that the methods analyzed, which use the concept of entropy, are adequate and efficient for evaluation of existing fluviometric networks, since they allow the reduction of eventual redundancies and at the same time, seek to maximize the information generated. It was possible to compare them and indicate the most appropriate method for the application within the national context, as well as indicate new methods for use thereof.