In this research, the Saudi demand model for imported Bananas was estimated, it determined by two explanatory variables, which are Bananas import prices and income per capita. Through unit root test by ADF, imports and values be integration from the degree of zero, while import prices, income per capita and values be integration from degree one, thus autoregressive and distributed lag ARDL methodology can be applied. The output of the R statistical program showed that the ARDL(2,3,3) model in its double-log form with a drift is the best model. Cointegration test was conducted using the boundary test and it was noted that there are long-term equilibrium relationships. The parameters of the model were significant at the level of significance of 10% and agree with the economic logic. The import price elasticity as well income per capita great than one, so the elasticity of Bananas imports to each of independent variable elasticity. By estimating the unrealistic error correction (UECM), the error bound coefficient was negative, significant, and shows that Bananas imports model in the short run could reach to the long run equilibrium in less than two years. The diagnostic tests showed that the model under study does not have econometrics problems, in addition, the model parameters were structurally stable during the study addition coefficients of the ARDL(2,3,3) model are structurally stable over the time interval under study. Finally, the model was used is to predict the Saudi's imports from Bananas during future period 2020 – 2029.