In response to the low efficiency and insufficient accuracy of traditional detection methods when facing micro defects, a micro defect detection method for electronic components on the basis of blending attention and Full-Scale Feature Enhancement Module (FFEM) is proposed. This method proposes a blending attention by combining spatial and channel domain attention, and incorporates a bidirectional feature pyramid structure to achieve multi-scale feature fusion, thereby improving the efficiency of feature extraction. The test results on the PCB and NEU-DET datasets showed that the Weighted Blending Attention Module (WBAM) achieved the average accuracy by about 10%, 15%, and 20% respectively at thresholds of (Mean Average Precision, mAP)@0.5, mAP@0.75, and mAP@0.9. The FFEM also significantly improved in precision and recall, with an average precision increase of 12% and a recall increase of 18%. In addition, the model that combined WBAM and FFEM further improves its performance indicators, with an F1-score increase of about 4%, accuracy increase of about 3%, and recall increase of about 5%. The research provides efficient and accurate new technological solutions for detecting micro defects in the electronic industry, which is of great significance for improving production quality and reducing production costs.