Estimation and prediction of temperature in Iraq using the multi-layered neural network model

The forecasting using the multi-layered neural network model is one of methods used recently in forecasting, especially climate forecasts for certain regions, because its accuracy which sometimes reaches levels close to real collected data. In this research, daily temperatures Iraq were predicted, by taking data from Iraqi Meteorological Authority (228) observations, represent Karbala Governorate year (2021), results autocorrelation and partial showed that temperature series governorate unstable, was confirmed conducting augmented Dickey Fuller test. analyzed two stages, it later shown estimation prediction even if time not stable, an indication rising increase during coming years. researcher concluded necessary pay attention vegetation cover conduct many predictive studies network.