Abstract: An online surface defect detection system for steel strips was designed to solve the problem of low efficiency in traditional manual methods. Firstly, the overall design scheme of the system, including the hardware structure, software structure and image acquisition system, was proposed. Then, the preprocessing and segmentation methods of images, the extraction and selection of features and the defect classification methods were studied. With the help of features extracted from the frequency domain image of defect region and defect classification based on artificial neural networks, the accuracy of the defect classification was improved. Finally, the system was tested by the samples of common defects and the experiment results verified the effectiveness of the proposed algorithms.