Provided in the present application are a track query method. The method comprises: acquiring query sets of descriptions comprising all query points; converting the descriptions of all of the query points into subject probability distributions which have position and time labels attached corresponding to all of the query points; searching for and matching semantic track data sets in a database for all of the query points on the basis of the subject probability distributions corresponding to all of the query points, and searching for candidate track sets of the query sets; calculating the distance between the query sets and all of the candidate tracks in the candidate track sets on the basis of the candidate track sets of the query sets, and sorting all of the candidate tracks in the candidate track sets of the query sets according to the length of the distances; and outputting the results to a user according to the sorted candidate track sets. Also provided in the present application are an electronic device and a storage medium. The present application may increase the accuracy of searching.