Abstract: To address issues such as fragmented process knowledge and low efficiency in welding planning for building steel structures, this study proposes a hybrid knowledge representation method combining ontology, frame, and production rule representations. Based on the characteristics of different knowledge types, a hierarchical knowledge model with a three-layer structure—"fact database, case database, and rule database"—is constructed. Ontology modeling is employed to structurally integrate factual knowledge, establishing a semantic foundation; Frame representation method is utilized to efficiently organize and retrieve case-based knowledge; Production rules are adopted to express and reason about rule-based knowledge. The three-layer knowledge base collaboratively enables semantic understanding, case matching, rule-based reasoning, and dynamic knowledge updating. Experimental results demonstrate that this knowledge model improves planning efficiency by 64.9% compared to traditional methods, significantly enhancing knowledge utilization and process planning efficiency.