---

Predictability of the exchange rate is one of the puzzles in international economics studies. During the past decades, it has been assumed that the economic models (especially monetary models) are the main instruments for the exchange rate forecasts. The poor performance of these models in forecasting exchange rate in addition to the weak support of empirical studies caused serious doubts arise in the usefulness of the monetary models. The purpose of this study is to forecast Iranian exchange rate using some monetary models and comparing the estimated results with Random Walk model. First, the considered monetary models have been estimated by VECM method for 1973-2008. Then, the exchange rate has been forecasted with the data during 2009-2011.The results show that comparing to other monetary models, Random Walk could forecast the exchange rate more accurately. In addition, compared to Frankel-Bilson, Frankel- Bilson with rational expectations and Frankel- Dornbusch monetary model with rational expectations, Frankel-Dornbusch monetary model has better forecasting.