Network traffic congestion is the major challenge in telecom service providersrnwhere usually they use a limited resource to deliver the services to their customers.rnThat leads to network performance and Quality of Services (QoS)rndegradation so that do not meet customer satisfaction. The Very Small AperturernTerminal (VSAT) network in Ethio Telecom delivers broadband servicesrnthrough satellite with a limited capacity. Therefore, this thesis aims tornstudy the VSAT network traffic patterns to propose the traffic forecastingrnmodel. That will be used as a solution to enhance the network resourcesrnbased on the prediction of the future traffic demand and as input for networkrnplanning and optimization works.rnIn this study, the VSAT data traffic recorded for one year from 01-Mar-2020rnto 28-Feb-2021 was collected. The dataset is used for data preprocessing,rnstatistical analysis, model training and testing. All the tasks are performedrnby Python software. In addition, existing time series forecasting methodsrnare selected from statistical and machine learning models that includes thernExponential Smoothing Methods (ESM), Autoregressive Integrated MovingrnAverages (ARIMA), Seasonal ARIMA (SARIMA), Artificial Neural Networkrn(ANN) variants Multilayer Perceptron (MLP), Recurrent Neural Networkrn(RNN) and Long Short Term Memory (LSTM). The forecasting accuracyrnmetric of Root Mean Square Error (RMSE), Mean Absolute Error (MAE)rnand Mean Absolute Percentage Error (MAPE) are used to evaluate thernforecasting performance of the models. These are applied to examine andrnchoose the model having the minimum forecasting errors. As a result, thernRNN model is identified the best model and improved the forecasting performancernby 44.94% than the Triple Exponential Smoothing (TES) model which is a variant of ESM. Therefore, the RNN model is proposed to EthiornTelecom for future network planning and optimization to VSAT networks.