Modeling of yarn and fiber properties has been a popular topic in the field of textile engineering in recent decades. The common method for fitting models has been to use classical regression analysis, based on the assumptions of data crispness and deterministic relations among variables. However, in modeling practical systems such as cotton spinning, the above assumptions may not hold true. Prediction is influential and we should therefore attempt to analyze the behavior and structure of such systems more realistically. In the present research, we investigate a procedure to provide a soft regression method for modeling the relationships between fiber properties, roving properties, and yarn count as independent variables and yarn properties as dependent (response) variable. We first selected the effective variables by multivariate test (mtest) and then considered fuzzy least squares regression for evaluating relationship between cotton yarn properties such as tensile, hairiness, unevenness and fiber properties that were measured by HVI system. We also used mean of capability index (MCI) to evaluate the goodness of fit of the fuzzy regression models. The results showed that the equations were significant at very good MCI levels.