In this work we present a procedure for the automatic classification of eclipsing binaries. The procedure is based on the data from a new catalogue of 6330 eclipsing variable stars, compiled by the authors and representing the largest list of eclipsing binaries classified from observations. The procedure allows the classification of a given system basing on a set of observational parameters even if the set is incomplete. Results of an application of the procedure to a test sample of a thousand classified systems and to a sample of about five thousand catalogued unclassified systems, as well as to stars from lists of eclipsing variables obtained as by‐products of microlensing surveys (OGLE, MACHO, ASAS‐3), are discussed. The classified eclipsing binaries can be used for determination of astrophysical parameters of their components, and the procedure can be also applied to coming data from future ground‐based and space observatories.The resulting data will be included in the Binary star database, BDB that p...