Tropical cyclones are extreme hydrometeorological events whose impact causes human, material and economic losses. The inaccuracies in the forecast of the trajectory of these phenomena often lead to inefficient decisions, such as unnecessary evacuation. This research proposes the combination of the three forecasting tools Numerical Tools for Hurricane Forecast (NTHF), Sistema de Pronóstico Inmediato (SisPI) and Sistema de Predicción Numérica Océano-Atmósfera (SPNOA) in the generation of ensemble forecasting systems, with the aim of improving the trajectory forecasts of tropical cyclones. Three variants were used for the construction of the time lagged ensemble sets, and for their evaluation the best track and historical errors (2016-2020) of the National Hurricane Center (NHC) were used. The ensembles lead to improved track forecasts for tropical cyclones. Position errors vary from case to case, but ensembles generally tend to be more accurate than independent forecasts. Compared to the historical errors of the NHC, the results obtained are promising because they are superior in some cases.