Railway marshalling yards are crucial for the disassembly and assembly of trains, and its network coverage is vital for rail transport efficiency. Traditional network optimization solutions struggle with base station interference and sensitive electromagnetic environments in this scenario. A wireless environment digital twin platform built based on ray tracing and its system architecture, optimization process and result visualization are introduced. Taking a marshalling yard along the Daqin Railway as a case study, the platform is used to simulate and optimize weak signal coverage. The simulation results show that compared with the initial base station deployment plan, the overall signal received quality is effectively improved after base station parameter optimization. Therefore, it can effectively assist in the design of base station parameters. This study provides theoretical support and technical accumulation for high-quality optimization of 5G-R networks for railway marshalling yard scenarios.