The transformer already connected to power system can experience saturation during the inrush transient of an incoming transformer. In the case of parallel energization, transformer can produce false differential currents that could cause relay mis-operation. In this paper sympathetic inrush phenomenon between transformers is analyzed using SIMPOW-STRI power system simulation program and laboratory transformers and a differential algorithm that is based on Artificial Neural Network (ANN) is proposed. Proposed network was trained using back-propagation algorithm. It is seen that the network is able to discriminate no-fault examples including sympathetic inrush condition from fault examples with high accuracy. Copyright (C) 2009 Praise Worthy Prize S.r.l. - All rights reserved.