Abstract: Strip steel is an important raw material in modern manufacturing,and its surface defects seriously affect the performance of the final product. In order to achieve accurate recognition and classification of surface defects in hot rolled strip steel,a method for identifying surface defects in strip steel based on feature weighted fusion was proposed.For 6 types of surface defects of hot rolled strip steel with small differences and large intra-class differences,machine learning algorithms were improved from three aspects of feature extraction,feature fusion and feature dimensionality reduction to obtain low dimensional and high information defect features.Features were trained through Support Vector Machine (SVM) to achieve defect classification.The results show that fused features have stronger representation ability than single feature,and the defect recognition rate of the method reaches 98.33%,indicating strong practicality.