Contingency ranking is performed to choose the contingencies that cause the worst effects on the system. Ranking by conventional techniques is time consuming and very tedious process. An alternative solution is the off line training and run time application of artificial neural networks. Therefore in this paper an Artificial Neural Network based approach is proposed for fast contingency ranking. A multi – layered ANN is trained using error Back Propagation Technique with the input data taken as the voltage, phase angles, active and reactive power values and output data as condition number obtained from off line load flow studies. The condition number reflects the severity of the contingency on the power system. This neural network structure is applied for contingency ranking of line outages in an IEEE 30 bus system and is found to produce almost accurate results for previously unseen line outages almost instantaneously. © Research India Publications.