This paper aimed to analyze the impact of oil price shocks, both negative and positive, on economic growth in Iran and Japan using an autoregressive distributed lag model. In doing so, a generalized autoregressive conditional heteroscedasticity GARCH (1, 1) model has been fitted to real oil price during 1980 Q1-2006 Q4. After capturing conditional variable and analysis of the existence ARCH effect on this variable, oil price shocks are then computed. Four single models are used to estimate the impact of oil price shocks on gross domestic product growth. According to the results, since the existing coefficients are not statistically significant, it can not be clearly decided whether the relation between these two variables in Iran’s economy is symmetric or asymmetric. In Japan, the estimated coefficients prove the existence of an asymmetric relation.