In our earlier work, we have proposed a cyberinfrastructure-based collaboration system for modular product design. One of the main components of the system is a design repository to which suppliers can upload the descriptions of their components using machine-readable, interface-based component description language, so that manufacturers can refer to the descriptions during product design phases. In this paper, we propose an efficient algorithmic approach based on a branch-and-bound (BnB) algorithm to support product design using the interface-based component descriptions stored in the design repository. This product design problem is categorized into a planning problem, whose complexity is known as non-deterministic polynomial-time (NP) hard. For performance evaluation, we compare the performance of the branch-and-bound algorithm with that of a depth-first search (DFS) algorithm, which is an exhaustive search method. This paper describes the details of the proposed branch-and-bound algorithm using a case study and experimental results are discussed.