The purpose of this chapter is twofold: (1) introducing of a semantic modeling mechanism, which is applied to achieve context-based knowledge fusion in a decision support system and (2) discovery of context-based knowledge fusion patterns. An approach to ontology-based resource modeling is proposed. The set of resources comprises sources of data/information/knowledge, problem solving resources and various actors. The knowledge fusion patterns are generalized with regard to two aspects: (1) preserving internal structures of multiple sources from which information/knowledge is fused within the ontological structure of context and preserving internal structure of the context itself, and (2) preserving autonomies of the multiple sources and the context. Six knowledge fusion patterns have been discovered. They are simple fusion, inferred fusion, instantiated fusion, adapted fusion, flat fusion, and historical fusion.