Abstract: The power transformer is one of key equipments in power systems. To evaluate the real time state of power transformers objectively, and analyze the risk caused by a fault, a risk assessment model based on fuzzy analytic hierarchy process method and support vector machine is presented. With the introduction of fuzzy AHP (FAHP), the loss extent of transformer failure is calculated by taking such four risk factors as equipment, environment, personal and grid security into account. By using support vector machine (SVM), the amount of health status is taken as an input, and transformer failure rate is used as an output, which can predict the transformer failure rate. In the end, the risk assessment of a regional transformer is conducted. The simulation results show that proposed method can improve the effectiveness and accuracy of risk assessment, and provide practical reference for risk management and maintenance decisions of transformer.