[en] The improvement of public transport in large cities is a fundamental factor for the quality of life. Poor transportation leads to an increased of greenhouse gases generation, hinders access to essential services and emphasizes the difference between social classes. One possible way to improve traffic in large cities is to encourage people to use public transport. By improving the quality of public transport systems and reducing tariffs, more people can use it as a means of getting around in urban centers. This article performs an analysis between different strategies (Neural Recurrent Network using LSTM and GRU, Convolutional Neural Network and ARIMA models) to model the variation of the occupancy rate (PTO) of the metropolitan buses in order to improve the planning and management of public transport. Results show that ARIMA models present better results to PTO forecasting and to describe the behavior of time series. This kind of approach can be used, in practice, to adjust the number of buses and population demand, for a given period.