Abstract: In order to complete the research of accurate identification of the use of electrical appliances, an electrical parameter measuring device that mainly incorporates the electrical parameter measuring circuit, the wireless communication unit and the power supply unit is designed. The electrical parameter measuring device is mainly applied to collect the working current data of a single electrical appliance and transmit it to the upper computer for processing. After the received data is processed by compound digital filter, the upper computer can generate training data-set for the purpose of electrical appliance recognition by a variety of permutations and combinations based on BP neural network through superposition principle. Experimental results show that the recognition of the usage of all learned electrical appliances can be completed with only a small number of samples; the recognition system has higher accuracy than traditional methods in a more complex situation where multiple electrical appliances are used at the same time.