This paper reports a refined model to predict the tensile behavior of rotor spun yarn from the knowledge of specimen specifications and yarn strain. Fundamental approaches used in modeling include a discrete fiber-modeling principle, an energy method, and a "shortest-path" hypothesis. The study is based on a coaxial-helix structural model, taking account of nonuniform yarn packing density. Most importantly, a changing-pitch system is introduced for the first time in yarn modeling. A pitch function is thus determined with the aid of a nonlinear regression method and a curve-fitting approach using images obtained from a tracer fiber technique. The model is also compared with experimental data, together with a theoretical analysis of fiber strain distribution, confirming that, due to the introduction of the changing-pitch system, the model is very successful in explaining the experimental observations.The study of yarn structure and the positioning of constituent fibers to form a continuous strand have attracted the attention of many textile researchers [4, 5j. However, most current yarn models are developed for general use, and little attention is paid to the difference between yams produced on various spinning systems. A typical example is that pitch, the yarn axial length within a whole twist, has always been identified by previous researchers as the reciprocal of twist level, a machine . data product. This is, of course, not appropriate when considering the inherent variations in different yam types. For instance, twist insertion in rotor spun yarn operates layer by layer. As a result, its twist magnitude from yam axis to yam surface cannot be a constant and thus cannot be obtained from machine twist data, as we often do with ring spun yarn. This phenomenon has never been included in previous models known to us and hence has become one of the subjects of this work. .A prominent feature of our work is that it adopts the energy method, which has been successfully used in modeling the mechanical deformation of textile fiber assemblies [!-3, 6, 7]. Moreover, the packing density of a yarn cross section is no longer considered uniform but a function of radial position. We also include Camaby's &dquo;shortest-path&dquo; hypothesis jt]―under tensile load, a yam will be separated into two regions, a jammed region and an unstrained region surrounding it. The jammed region implies that the packing density of this region has reached a prescribed maximum value. Tensile strain can only be accumulated within fibers in this region because fibers outside can readily move inward or outward to avoid being strained. Based on these considerations, we attempt here to postulate the phenomena likely to occur, to construct a model that represents them, and to test the validity of that model against experimental observations.