Predicting the future and figuring out what to do with it has always been the real business. Prediction aims to assist users in making decisions relating to the exercise of economic activity. From time series data we predict USD-IDR exchange rate by using Average-Based Fuzzy Time Series Markov Chain Model. This method includes four main concepts, Average-based to determine the effective length interval, the concept of Fuzziness to classify variables, Time Series to observe the exchange rate movement during a particular period and Markov Chain used in prediction process by conduct transition probability matrix. Result using this method will be compared with previous method (Fuzzy Time Series Markov Chain), and a web based application will bulild so that the prediction process can be more efficient, thorough and practical.