One of the most important and effectiveness of macroeconomics variables is prediction of future exchange rate trend which heavily considered by economic scholars. Its changes affects different parts of economic, thus it is necessary to model it to provide more suitable economic advising. In order to do that, in this paper we have used SARIMA, ARCH and GARCH models to simulate the time series trends of exchange rate in Iranian non-official market. The results show that GARCH (Generalized Autoregressive Conditional Heteroskedastistiy) provides better and more acceptable outputs than SARIMA.