The rapid progress in the industry increases the number of plant, and a myriad of operation and maintenance manuals is required for the plant safety activities. Moreover, major challenge in safety prevention and maintenance is that manual search and retrieval requires extensive labor-intensive and timeconsuming process incurred by high degree of semantic ambiguity and linguistic complexities among the various operation and maintenance manuals. In this paper, we propose a novel approach on plant operation and maintenance guideline using semantic processing and clustering techniques for the plant safety activities. Specifically, the proposed method can provide the engineers with the more enhanced guidelines by information propagation.