Abstract: In the automatic wire-traction micromanipulation system for the hollow cup coil, the coil's geometric center is difficult to coincide with the rotation center of the rotary stage, thus, the eccentricity occurs during the rotation, which leads to a failure of the wire-traction. Therefore, a method to detect and compensate the coil's eccentricity is proposed in this paper. Based on the analysis of the eccentricity sources, the eccentricity mathematical models are established, a bias measurement method combining the eccentricity models with the image feature extraction algorithm is designed, and a correction method based on hardware and compensation models is proposed. The accuracy of the models and the effectiveness of eccentricity compensation were verified by experiments. Relying on the evaluation of root mean square error (RMSE), the prediction for eccentricity based on models and the feature extraction algorithm both had a high accuracy. The maximum residual error after compensation was less than 4μm and the average compensation amplitude was about 98.6%, and the wire-traction experiment showed the excellent performance of the proposed method. The detection and compensation method for eccentricity proposed in this paper has the advantages of high accuracy, low complexity, high efficiency, real-time and embedded, and is more suitable for a wider range of application in the field of micromanipulation and micro-assembly.