To solve the manufacturing resource description problem and the service matching problem in cloud manufacturing, a general semantic description framework for manufacturing resources which considers the user's personalized needs and the heterogeneity of manufacturing resources was proposed. The difficulties of the cloud manufacturing resource description and the service matching were analyzed. By using the features of standardized description, the manufacturing resource description problems were transformed into Web semantic description problems. Combined with four similarity algorithms and the design of threshold and weight, a service matching strategy with filter was proposed. The experimental results verified the feasibility and effectiveness of the proposed strategy and algorithm. The application of the four similarity matching algorithms can effectively reduce the matching scale and improve the service search efficiency and accuracy.