Time series forecasting is a field which paid a considerable attention of the research community. At present, in the field of machine learning, there are a lot of studies using evolutionary algorithms trained artificial neural networks to construct the model of time series forecast in general, and foreign currency exchange rates forecast, in particular. There are two issues to consider: First, the global convergence of the extreme solutions using evolutionary algorithms
Second, determine the optimal weight of the network. In this paper, the authors propose an algorithm designed multi-objective evolutionary (NSGA-II) trained neural network and application problems in forecasting exchange rates. Two objectives of the selected model include: MSE and DIY. The model proposed in the four experimental data sets monetary and compare with single-objective method which the research community did. Through MSE indicators from experimental results show that the model that the authors proposed to forecast better results some simple methods existing goals.