Abstractâ€”This paper discusses the problem of modeling the computer network bandwidth usage forecasting using averages based-fuzzy time series (FTS) are developed with a combination of Markov chains. This forecasting model development aims to get better forecasting results, especially regarding the accuracy of forecasting. Development of methods is done by inducing Markov chain method with mathematical rules and applied at a particular stage. The development model is applied to forecast traffic data bandwidth usage on a computer network. The results show that the average based FTS forecasting model with the addition of a Markov chain has a value of forecasting accuracy better than average based FTS method, with the percentage increase in average accuracy of 41.590%, if the prediction error calculation using the average difference MSE and amounted to 30.348%, if the forecasting error calculation using the average difference MAPE. With the expected results of this innovation can be applied to the management traffic bandwidth on the network computer.Keywordsâ€” Method development, average based FTS, Markov chain, bandwidth usage.