As the industrial digitalization has been recently leaping forward, high-efficiency production processes emerge, and assembly integration technologies are being progressively developed. For high-tech industries (e.g., aerospace manufacturing), however, unprecedented challenges are imposed on the conventional assembly manufacturing process technology. To be specific, the shallow digitizing and intelligent technologies fail to satisfy the rising requirements for efficiency, pass rates and cost; in particular, the way to down-regulate the unqualified rate and assembly cost of parts production remains unclear. In the present study, an intelligent matching assembly system was designed for reapplying out-of-tolerance parts. In addition, the concept of the mating base is proposed. Next, the fit relation between parts is quantified by employing the functional expression of dimension data, and such a relation is characterized as a function constraint with a variable of dimension. Moreover, the designed intelligent matching system was introduced into the existing assembly process to reemploy a number of unqualified parts and determine the matching parts most suitable for the original unqualified parts. On that basis, the most suitable parts are enabled to be involved in the assembly activities, as an attempt to satisfy the design requirements. To efficiently find the optimal matching scheme in the database of the unqualified parts and most effectively employ the unqualified parts, a set of optimal matching algorithms are proposed, thereby significantly down-regulating the production cost. Lastly, Monte Carlo method was adopted to simulate the actual production data, and the optimal matching scheme determined by the intelligent matching system was analyzed and assessed, which verified the effectiveness and practicability of the system.