Abstract: In view of defect image of edge blurring and low contrast, the traditional fuzzy clustering method is easy to cause the target clustering errors and hinder the extraction accuracies of defect feature parameters, which lead to the defect misclassification and reduce the defect recognition rate. Therefore, a novel image defect recognition algorithm based on FLICM and geometric feature is presented in this paper. Firstly, the FLICM model is used to segment the defect image to acquire the image defect areas. Secondly, the multiple geometric features of image defect are extracted. Finally, the parameter variable combinations of geometric features are designed to recognize the defect target. The experimental results show that our model improves the recognition rates of the shrinkage, inclusion, crack, delamination, channel and blowhole defects.