In Manufacturing Grid (MG) system, there are primarily two kinds of users: resource service consumer and resource service provider. For a resource service request task, the MG system should search the services which are qualified for the requirements of consumer and then choose the optimal one to execute it. In according to the distributed system structure of MG and node routing model influenced by several factors (e.g., time minimization, cost minimization, bandwidth minimization), the searching technology based on web service technology and ant colony optimization (ACO) algorithm has been proposed in this paper. The basic evaluation models and formulations are described, and then the algorithms are presented to minimize the expenses. The experimental results also show that the proposed method is useful in solving the searching problems in MG system.