Aiming at reducing the frequency of the electrical shock faults in low voltage power system, approximate entropy (ApEn) based electrical shock feature method is proposed. Firstly, it samples a total residual current from a single-channel and slides a time window continuously to calculate the ApEn in every target period. Then, it determines whether the electrical shock fault happens by identifying the varying tendency of ApEn. Considering the redundant operation in the present algorithm, it proposes an improved algorithm to reduce the time consumption of electrical shock feature detection. Finally, physical experiment about IEC body impedance model, experimental rabbits and sprays is made and verified by using the data. The result shows that the method is of noise robustness, and also independent of amplitude of the total residual current and the way of electric shock. Moreover, the proposed algorithm shortens the detection time of electric shock feature, meeting the need of engineering application. This work is supported by Science and Technology Department of Sichuan Province, China (No. 2015RZ0055).