The invention relates to the technical field of maintenance of rail traffic, in particular to a method for predicting SOE of a rail traffic lithium battery through large data. The method comprises thesteps of a data preparation step, a data arrangement step, a data characterization step, a target determination step, a data calculation step, a training verification step and an algorithm evaluationstep; hidden noise data is found through special cleaning means, so that the effects of good cleaning effect, high accuracy and the like are realized; in addition, model training and evaluation are carried out, different algorithms are selected for matching verification and are issued by using different models of machine learning through data import, so that a structured product is formed, and the prediction accuracy of the models is constantly improved along with time accumulation and data enrichment.