Abstract: Existing virtual construction methods suffer from uneditable workflows, non-reusable construction knowledge, and low modeling efficiency. To support rapid visual rendering of diverse construction procedures and techniques, this paper proposed a parametric template-based method for railway virtual construction and knowledge reuse. The paper mapped construction sequences to key frames, established digital twin models for construction workflows, and built a sub-joint attachment linkage model for construction machinery. It extracted virtual construction knowledge and parameterized such knowledge to form construction method knowledge templates. The paper adopted multi-track information fusion based on time series to realize visual representation of virtual construction. Experimental results show that the proposed method realizes knowledge reuse at the construction procedure and technique levels and improves the production efficiency of virtual construction by more than 80%. This method provides a multi-level knowledge reuse scheme for railway virtual construction practice.