the conditions they are produced and combined them with special characteristics and can be optimized to create. In this project, several samples of different yarn polyester viscose and cotton material with different counts by Intermingling jet are mixed together and then some physical characteristics and quality of yarn produced is measured. Properties measured in this study include cv%, Thin , Thick , Nep , hair and physical properties including strength and elongation are. Then with the sample data And using Perceptron neural network algorithm backward error, yarns of different properties have been predicted. First of all the software Excel data obtained were classified according to the seven parameters and the physical quality and the above mentioned has been drawn in separate graphs and the detailed results are discussed. More to obtain the best yarn produced from above characteristics has been helping of the genetic algorithm. It seems that the best prediction of cotton yarn filament by the hybrid system with count of 7 and then polyester viscose yarn hybrid count is 17.