Despite the considerable progresses in computational hydraulics the calibration of hydrological models still requires a meticulous and laborious effort. However, the development of model calibration supporting tools has facilitated that work and improved the obtained results. The objective of this paper, developed under a R&D project, is to evaluate the performance achieved by using different calibration techniques for the rainfall-runoff Sacramento model of the Itabapoana-MG river basin. Bom Jesus do Itabapoana, a flood prone city, is located about 35 km downstream of Rosal hydropower plant. The limited capacity of Rosal reservoir implies that it cannot be used to mitigate flooding at downstream cities. However, flood forecasting can be used to avoid potential downstream damages and support decision making to defend and mitigate the undesirable effects. There should be a balance between the models performance and their computational times. It is important consider sufficiently precise results for sound decision-making outcomes, as well as results fast enough to be used in an operational decision support system environment. SOBEK software was chosen for prediction of flood events and Sacramento hydrological model was selected as rainfall- runoff model. SOBEK does not offer a module for automatic calibration of their hydrological models. Thus, for the quantification of Sacramento parameters, it is possible to make use of external algorithms for automatic calibration based on global optimization techniques. This paper assesses the eficiency of using results of automatic calibration using the Rainfall Runoff Library (RRL) within SOBEK hydrological module. In general, the results for RRL were better than the ones reached by SOBEK. Despite differences, the use of RRL as a tool to accelerate the parameters determination process for the Sacramento model has obtained satisfactory results.