Polyester fibers were tested by a method combining size exclusion liquid chromatography and reversed phase chromatography techniques. Results from several cluster methods were used to identify the interior structure of measured data and the existence of natural clusters. As to their peak structure and area distribution, chromatograms for the six types of polyester fibers differed only insignificantly. Therefore, possible use of multivariate pattern recognition techniques was checked for polyester fiber identification based on a data set obtained through HPLC analysis. The application of cluster methods (dendogram, display mapping, and potential density method) enabled the classification of fibers into different manufacturers. It is pointed out that the chromatographic procedure should be viewed as a method for use in addition to other well-established methods associated with forensic fiber analysis. 37 references, 4 tables, and 5 figures (Author abstract modified)