In the weaving process, especially in the loom shed process, yarn breakage is frequently found to be a critical problem since it results machine downtime. This study aims to develop an optimal set of process parameter values by taking speed, air pressure, warp tension, backrest height, and relative humidity. Signal to noise ratio, optimal parameter combination, and ANOVA are determined using a Taguchi designed experiment. All parameters are found significant and the yarn breakage rate is reduced from 87 to 32 per shift. Possible optimal regions and a predictable regression model is built using the response surface method. The regions are found capable by reducing the breakage rate to less than 32 per shift. Moreover, a genetic algorithm is employed using Matlab2014a so that the lowest possible yarn breakage rate is recorded, for which a model is identified using best-fit regression analysis. It can, therefore, reduce the yarn breakage rate from 87 to 8 per shift. Thus, this finding is determined to be significant for various textile weaving industries.