A automatic recognition of wear debris extracted from mechanical equipment oil based on Fuzzy support vector machines was proposed.The wear debris image extracted from mechanical equipment oil was segmented by Kmeans clustering algorithm to extract the feature parameters as their quantitative characteristics,including shape feature,edge detail and surface texture feature.The parameters that best reflect the characteristics of the wear debris to be identified were selected as input vectors for each twoclass classifier.Based on the decision tree and oneagainstrest method,a new hierarchical multiclass classification model of fuzzy support vector machines was established.The particle swarm optimization was used to optimize parameters of the model during the training process.The proposed model was applied to identify wear particles extracted from lubricant and hydraulic fluid of a rotary drilling rig,the recognition rate was up to 90%.The proposed model has advantages of simple structure and good classification accuracy,and has a great engineering value in the field of wear debris recognition.