Among many applied techniques for overhead transmission line fault recognition, the artificial neural networks-aided schemes have demonstrated superior efficacy. This paper presents two methods based on time domain measurements at the local end of the transmission line to identify the single circuit line fault type for low impedance faults. The first deals with time samples of the current and voltage waveforms of the three phases during the fault. The second uses the measured mms values of currents and voltages. The latter can also determine the fault location on the line. It is also extended to produce the faulty circuit and fault type of double circuit lines for low impedance faults. Furthermore, the paper proposes a new algorithm for diagnosing the single line high impedance fault. A modified version of the algorithm is also developed to obtain a good representative feature vector with the least possible elements. The modification has enhanced diagnosing capabilities and can classify the high impedance faults. This enables the fast and accurate recognition of the faults that is a necessary request for the digital relaying system. The design and training of the decision making ANN for each approach are described. The studied techniques are tested with many cases to assess their diagnostic capabilities. Besides, the performance of the competitive methods is compared.