Abstract: In order to solve the problem of selecting knowledge resources in the cloud manufacturing environment， a selection model is presented based on the capacity of knowledge service from the perspective of knowledge resources suppliers’ service capacity. Through establishing the system for evaluating the capacity of knowledge service， the variable precision roughness set algorithm is utilized to calculate the weight of each evaluation index， and the sensitivity coefficient is used to revise users’ preference in actual cases. All the candidates’ capacities of knowledge service are obtained， based on which to guide the cloud platform to select the best supplier for users. Finally， examples are given to prove that the model can select the best supplier for users.

Key words: cloud manufacturing, knowledge resource, capacity of knowledge service, preference revision, variable precision roughness set