Wheat is one of important cereals for human nutrition that its importance is mostly related to physical and chemical traits of gluten in wheat seed. The impact of mutation on traits can be determined by investigation of rheological behavior to study the physicochemical properties of bread dough in studied genotypes. Genetic variation was created by gamma ray from the cobalt-60 source. These genotypes were selected by specific primers related to LMW-glutenin subunits and change in Dx2+Dy12 alleles to Dx5+Dy10 ones. Chemical and Farinograph experiments were tested on seeds. In this study, sigmoid transfer function was used for assessment of factors in three layers by the model of feed-forward neural network with training method of levenberg-marquardt algorithm. Three chemical traits of Zeleny number, the hardness, wet gluten, and protein content in Roshan3 line increased significantly compared to the control. Studying Farinograph traits In Roshan3 line indicated significant increase in water absorption percentage, dough stability and valorimeter value but dough softening after 10 and 20 minutes reduced significantly compared to the control. In the artificial neural network carried out based on levenberg-marquardt algorithm for chemical and Farinograph traits protein content, bread volume, Farinograph quality number and E10 had the greatest impact on neural network model. These results indicate that mutation is able to change the qualitative characteristics.