The purpose of this study is to conduct an experimental investigation into the ability of multivariate economic models. We will use the linear lagging autoregressive models approach and the artificial neural network approach and use fundamental macroeconomic variables, including productivity differences between Algeria and trading partners, oil price, degree of trade openness, and government expenditures, to examine whether the artificial neural network models have the ability to better predict the effective of the real Algerian dinar exchange rate.The results showed that the multivariate artificial neural network models are better and more efficient in the process of predicting the exchange rate of the Algerian dinar.