This study was carried out in the province of Esfahan in Iran in order to model field emissions of wheat production, using artificial neural networks (ANNs). Data were collected from 260 wheat farms in Fereydonshahr city with face to face questionnaire method. Life cycle assessment (LCA) methodology was developed to assess all the environmental impacts associated with wheat cultivation in the studied region. Global warming potential (GWP), eutrophication potential (EP), human toxicity potential (HTP), terrestrial ecotoxicity potential (TEP), oxidant formation potential (OFP) and acidification potential (AP) were chosen as target outputs. System boundary and functional unite were selected farm gate and one ton of wheat grain. All input energies and farm size were selected as inputs and six impact categories were chosen as outputs of the model. To find the best topology, several ANN models with different number of hidden layers and neurons in each layer were developed. Subsequently, we applied different activation functions in each hidden layer to assess the best performance with highest coefficient of determination (R2), lowest root mean square error (RMSE) and mean absolute error (MAE). Accordingly, ANN model with 12-6-6-6 structure showed the best performance. RMSE for GWP, HTP, EP, OFP, AP and TEP were 45.82, 6.22, 7.47, 0.96, 0.28 and 0.09, respectively. Also, MAEs for this model were 14.9, 0.77, 1.5, 0.02, 0.14 and 0.02 for GWP, HTP, EP, OFP, AP and TEP.