The practice innovation in cloud manufacturing and related mechanism have been widely investigated in the fields of supply chain management and manufacturing management for a few years. The fragmented resource matching problem on a cloud manufacturing platform is studied to find a way to improve the efficiency of enterprise resource allocation. With multi-attribute evaluation under the prospect theory, satisfaction degree values and corresponding preferences between sellers and buyers are calculted. And based on agents' preference information, a heuristic algorithm is designed to obtain a stable matching result for the many-to-many matching problem with matching quantities. It is demonstrated that the matching scheme obtained is a Pareto-optimal solution for the agents on a dominating side. Finally, a sample application is given to illustrated the feasibility and effectiveness of the whole matching mechanism.