The aim of a decision support system (DSS) is to enhance the course of decision-making by furnishing a decision episode and support during the decision-making and enforcing procedure. Inspired by management information systems, the DSS assists policy makers to reach a decision by man-machine interaction through data, models and knowledge. Knowledge fusion is an effective means of enhancing the efficiency of a DSS. Thus, this paper presents a three-layer, six-step knowledge fusion pattern from the perspective of interaction and the three procedures used to arrive at a decision. For better understanding, an evaluation of the knowledge state with knowledge objects in the fusion pattern is also presented. Furthermore, we briefly examine a case study of a Fire Rescue DSS(FRDSS).