The large–scale industrial production of fiber optic cables is a multi–stage, complex process that requires the use of raw materials and semi–finished products at various stages of their manufacturing. One of the key problems in cable production is tube extrusion. Fiber is used for its production. Depending on customer orders, the production process is carried out in appropriate stages. Minimizing production costs necessitates time–consuming planning work to select the proper capacity of fiber spools so that the amount of waste generated is low or zero. In this work, an intelligent algorithm for selecting fiber optic strands based on warehouse stock analysis and declared order segments is proposed. Due to the strongly NP–hard nature of the problem, algebraic–logical meta-modeling technology was applied. A model was constructed that enables the search for feasible solutions to the spool matching problem within a finite time, meeting technological requirements and minimizing the amount of fiber scraps as well as retooling resulting from the generated production scenario. The simulation part of the work proposes the use of a parallel allocation mechanism of dedicated computational molecules, enabling the realization of this task in a parallel environment, thereby reducing the solution search time.