When the overhead crane is in normal operation, the wheel rim and the overhead crane steel rail should maintain a certain gap (20mm-30mm). In the production practice, the problem of overhead crane steel rail gnawing has always been a problem that affects its normal operation. It reduces the working efficiency of the equipment, increases the maintenance workload and operating cost, and leads to safety hazards, causing adult or equipment accidents. Therefore, it is very important to detect the status of the overhead crane steel rail in time and repair or replace it in advance to prevent a larger accident. At present, there are mainly two methods for judging the overhead crane steel rail, namely the experience method and the wheel detection method. The empirical method is mainly based on the phenomenon caused by the previous overhead crane steel rail to judge whether the current overhead crane rail is broken, such as obvious friction marks on the side of the overhead crane steel rail, the running resistance of the crane is increased, etc. The accuracy of this method is poor, and the wheel detection method is adopted. The horizontal deviation, vertical deviation and span of the crane wheel are measured to determine the track distance and the degree of curvature of the track, and the measurement accuracy is not high. Based on the principle of laser triangulation, this paper proposes a new crane track detection method and develops a corresponding detection system. The detection system mainly includes three parts, namely total station, orbit detection robot and reconstruction computing system. The orbit detection robot carries the corner prism along the track, the distance from the prism to the side of the track remains unchanged, and the total station tracks and records the position of the prism in real time. Through the data processing, the top line of the top surface of the two rails can be reconstructed, and the straightness of the crane track, the parallelism of the two rails, the track span and other parameters are calculated. The system is designed according to the cross-sectional dimensions of common rails and is suitable for crane rails with a width of 70mm-130mm and heavy rails with widths of 68mm and 70mm. The measurement accuracy is +/-2mm. The detection speed is 0.2-0.3 m/s, so the detection efficiency can be greatly improved.