This study utilizes the estimation of aggregate import demand under foreign exchange constraints in Ethiopia using annual time series data from 1985 to 2020. Regression analysis was carried out using the nonlinear autoregressive distributed lag (NARDL) approach to examine the impact of the accumulation of foreign exchange reserves on aggregate import demand in Ethiopia. The estimation results indicate that, in the long run, all the variables i.e., foreign exchange reserve, the relative price of import, real income, volatility of exchange rate, money supply, and policy dummy significantly determined the behavior of aggregate imports over the reference period. Findings also show that in the long-run foreign exchange reserve, real income, and exchange rate affect positively the demand for imports in Ethiopia. While a positive shock in relative import price and money supply negatively affect import demand in Ethiopia. Thus, the price and income elasticity estimates have correct signs and are statistically significant. The variables included in the model strongly affect import demand in both the short and long run. Accordingly, policy makers aiming significantly influence import demand through effective management of those variables as it strongly affects import volume. Keywords : NARDL, Import demand, and foreign exchange reserve