This paper describes the application of artificial neural network (ANN) techniques for protecting small power transformer 2 kVA (Terco type). ANN network trained according to the primary and secondary currents data under NN tool, this network performs as a function of differential protection relay. Symmetrical and unsymmetrical faults are analyzed using Matlab environment. ANN network senses the difference in the internal current of both transformer sides and sending a trip to the circuit breaker (CB) at moment of fault occurrence. All the voltages and the currents waveforms affected with the fault and the response time increased according to this technique. Finally, the trip signal and the quick disconnect time were improved according to ANN technique.