The details of an infrared technique standardized to analyze cotton–polyester blends of different compositions are reported. It is observed that particle size plays a very important role in the peak intensity of the 1725 cm−1 band (CO stretching) in polyester, the absorption band recommended for analysis of cotton–polyester blends. However, the product of the band width and peak intensity is found to be dependent on particle size to a lesser extent and hence is a more reliable parameter in the measurement of percentage composition. It has further been shown that the method of sample preparation can affect the band width and hence the results. A suitable sample preparation procedure is, therefore, described whereby an accuracy of ±3% in the estimation of polyester content in a given blend can be achieved.