This study uses the early warning system approach to predict banking crisis in Iran during 1990Q1-2013Q4. To achieve the goal, Money Market Pressure index approach will be used by Markov Switching Model. The results indicate that, although, based on governmental supporting, the banking section in Iran has never encountered the phenomena such as bank run and bankruptcy, but it has also experienced banking crisis. Likewise, the assessment of probit model suggests that some indexes are leading banking crisis probability. These indicators include the variables of real exchange rate growth, the growth rate of credit endowed to private sector, real GDP, housing price, and real interest rate. Furthermore, the measures of expectation-prediction represent that the model developed has considerable potential to predict in sample banking crisis. Also, this model is unsuccessful in the prediction of the crisis in only 12 percent, but capable of predicting crisis in 77 percent of cases, where the crisis has occurred with probability of more than 40 percent.