Abstract: A model and relative solving steps for power transformer fault diagnosis are proposed on the basis of Euclidean Clustering（EC） and Support Vector Machine（SVM） theory in this paper.Some key dissolved gases are selected as the inputs of the diagnosis model and data preprocessing is applied for these gases.According to thirteen fault types，a new model based on EC diagnosis model and SVM models is constructed.By comparing with the other methods，the proposed model reduces the error ratio，and recognition results show that this model is effective.