The machine made carpet industry is one of the main and most famous industries in Iran and especially in the city of Yazd. However there is little information about customer preferences for different attributes of this product. In this article we tried to estimate the relative importance of the main attributes affecting customer desire for purchasing machine made carpet and the utility values for the different levels of each one by means of conjoint analysis. In addition to this, we created customer segments with similar preference structures using cluster analysis. Six attributes have been considered in this paper: design, color, number of colors, density, primary material and brand. Twenty seven profiles by combining different levels of these attributes using fractional factorial design approach have been created. These profiles were evaluated by 380 customers in the city of Yazd. Results have shown that design of carpet is the most important attribute for the choice of carpet. Color, primary material, brand, density and number of colors are the next priorities for customers respectively. Also cluster analysis identified five clusters of customers with similar preferences.