Series arc fault caused by aging damage of line insulation layer and poor electrical contact seriously threatens the power safety of low voltage distribution system. It is difficult to detect and extinguish series arc fault for its characteristics of small current, high temperature and strong concealment. Because of above reasons, a method based on wavelet feature and deep learning is proposed for detecting series arc fault. Firstly, series arc fault experimental platform was built to collect the current signals under typical resistive load, inductive load and resistiveinductive load. Secondly, after transformed by wavelet transform, collected signals were decomposed to construct training sets and test sets. Finally, the arc fault was identified by the improved AlexNet model, and the test results were output. The experimental results show that the accuracy of this method for serial arc fault identification is almost 9558%, about 1058% higher than using AlexNet model.