A steel surface defect detection algorithm based on improved YOLOv7 is proposed to address the high false detection rate and missed detection rate in steel surface defect detection. In this algorithm , the ConvNeXt-CBS module is introduced to enhance the feature extraction capability of the network. Additionally , the MPCS module is constructed based on the SimAM attention mechanism to increase the network ’s focus on small defect targets. Finally , the C3 module is introduced in the model to improve network stability. Experimental results show that on the NEU - DET dataset , the detection accuracy of this algorithm reaches 80. 2% , which is 3. 9% higher than the YOLOv7 algo- rithm. Compared to previous steel surface defect detection algorithms , this algorithm achieves higher detection accura- cy and faster detection speed , making it highly suitable for industrial applications.