Abstract: The article develops the main factors affecting the quality of fibers and seeds and their interrelationships, as well as mathematical regression equations. It has been shown that the separation of fiber from seed in the process of sawdust fiber separation depends on many factors and their interaction and can be studied using modern mathematical methods of experimental planning that allow to find a solution to the problem. Acceleration of the speed of the raw material roller by means of a seed extractor and a new optimal profile of the working chamber were found. Through the planned experiment, the most important factors influencing the quality indicators of fibers and seeds and their interrelated effects were studied. The main factors and their levels of variability were determined by the output parameters during the selected experiment. The randomly ordered planning matrix of the experiment and the inequality of variation are presented. The adequacy of the obtained equations is shown by the closeness of the experimental results with the mathematical calculation models. Thus, a system of equations was created that allows to find the optimal parameters of the auxiliary seed extractor in the working chamber of the saw fiber separator.