The invention relates to an intelligent distribution network fault recognition method. The method includes the steps of firstly, scheduling switch accident tripping information of 110 kV or lower in an SCADA system, fault report records of a production management system (PMS) and distribution network fault data in a distribution network automated DMS by having access to power grids of all power supply companies; secondly, collecting telecommand displacement warning, secondary telecommand warning, the telecommand SOE, telemetering sample data and remote control operation record information in a local power supply jurisdiction, then sending the telecommand displacement warning, the secondary telecommand warning, the telecommand SOE, the telemetering sample data and the remote control operation record information to a distribution network fault information statistical analysis database, and forming TCM events by comparing and matching the telecommand displacement warning, the secondary telecommand warning, the telecommand SOE, the telemetering sample data and the remote control operation record information with tripping information reported by a fault repair management system TCM; thirdly, forming a highly-suspicious advance unaudited event set by applying fault judgment logic, comprehensively analyzing distribution network fault data on this basis, and obtaining effective fault tripping statistical data by analyzing fault causes. Compared with the prior art, the method has the advantage that the reliability and safety of the distribution network power supply are improved.