With regard to the role of fiscal policy in reducing the financial crises, determining the magnitude of the fiscal policy multiplier after the 2007-2008 global crisis became one of the most challenging issues in the field of fiscal policy. In general, the magnitude of the fiscal policy multiplier is estimated larger than one according to the Keynesian viewpoint and smaller than one based on neoclassical viewpoint. The difference in the magnitude of the multiplier comes from the fact that economists believe that the fiscal policy multiplier is influenced by the degree of economic openness, the exchange rate regime, the way monetary policy is applied and the business cycle. Differences about the magnitude of the fiscal policy multiplier among schools of economic thought are evident in both theoretical and empirical dimensions. In this regard, this paper tries to estimate fiscal policy multiplier using structural vector autoregressive model (Blanchard and Peroti, 2002 method and Markov switching approach (Hall, 2009) using the seasonal data for Iran during the period (1990: 1-2017: 3). The results of the structural vector autoregressive model showed that the instantaneous multiplier, 10-quarter cumulative multiplier, and the 20-quarter cumulative multiplier of the government expenditure were equal to 0. 281, 0. 304, and 0. 445, respectively. In addition, the corresponding multipliers for taxes were-0. 079,-0. 107 and-0. 171, respectively. Since the fiscal policy multiplier varies based on the economic conditions, the results of the nonlinear Markov switching model showed that the government expenditure multiplier during the recession is 0. 828 and it is larger than the same coefficient during the boom period (0. 108), on one hand. On the other hand, the tax multiplier during the boom period (-0. 194) is larger than its value during the period of recession (-0. 092 ).