Abstract: The Railway Snow Depth Monitoring System plays an important role in ensuring train safety in the severe cold areas. In terms of the distortion of snow depth monitoring data and the possible false alarm caused by the anomalous data of the system, the article analyzed the initial snow depth monitoring data of a railway line, sorted the anomalous data into different types, made data filtering rule, proposed a data filtering algorithm for Railway Snow Depth Monitoring System, and took a simulation testing. The simulation testing result showed that the algorithm could effectively filter out anomalous data and eliminate data jitter, be with significantly applying value.