For the real-time and objective measurement of the risk caused by on-line power transformer, a power transformer risk assessment method was put forward based on Fuzzy Analytic Hierarchy Process (FAHP) and Artificial Neural Network (ANN) from the perspective of the accuracy and quickness of risk assessment. An analytic hierarchy process model for transformer risk assessment was built by analyzing risk factors affecting transformer risk level. In this model, a risk index system was established and index weights were quantified. The weight relation of each risk factor in transformer risk calculation was analyzed by applying fuzzy consistency judgment matrix. It took major risk factors as the input of the neutral network by making use of adaptive ability and nonlinear mapping ability of the ANN, thus realizing the intelligent quantitative assessment of transformer risks. Simulation results have shown that this method increases the speed and result accuracy of risk assessment and can provide feasible decision basis for transformer risk management and maintenance decisions.