Abstract: Objective: Due to the limited working time, numerous inspection tasks, extensive inspection coverage, and long distance involved in track inspection work, the existing manual inspection methods are no longer sufficient to meet the requirements. Therefore, there is an imperative need for developing new means to enhance inspection efficiency. Method: The highspeed intelligent inspection system for tracks is installed on metro trains, utilizing 5G (fifth generation) mobile communications technology for rapid image transmission. Deep learning algorithms are employed to achieve realtime intelligent identification of track surface abrasions or detachment, corrugation, abnormal light bands, weld seams, missing fasteners or bolts, misaligned fasteners, broken fasteners, track plate cracks, and foreign objects on the track. Result & Conclusion: The trial results from the pass three months on Nanchang Metro Line 1 indicate that the system can detect diseases such as damage on rail top surface, fastener defects, track plate deficiencies, and foreign objects on the track. The recognition rate is not less than 95%, with a false alarm rate of not more than 5%, demonstrating high recognition rate and good repeatability. At the same time, it does not occupy the track inspection window time, enabling daily inspections and transforming conventional track inspection methods. With the possibility of daily allweather inspection work of track surface conditions, incalculable hidden economic and social benefits can be generated.