The dynamics of the drafting process for cotton sliver is investigated with regard to its principal role in the synthesis and optimization of automatic control systems for reducing sliver irregularities. The approach is formal: the models refer to spectral filters, and the parameters of the respective transfer functions are determined by ad ditional signals—the speeds of the drafter rollers. Considering some trivial initial con ditions as given and using integral transformations of known initial equations, math ematical models in the form of transfer functions are obtained for a double-zone drafter with fibers of equal or different lengths. The drafter is presented as an object with four inputs and one output. The models are also presented in digital form as discrete transfer functions. To estimate model parameters, on-site experiments with harmonic test signals follow standard identification procedures. Time series with values of the object main signals have been registered synchronously using a digital data acquisition system with minimal sampling period. A specially designed software pack age is used to simulate process dynamics and the parts of the control system.