In order to better formulate the load plan of microgrid, encourage the users participation in demand response, and improve the movement and regulation capability of the power grid, this paper proposes a customer-side load demand prediction method based on the discrete wavelet transform (DWT)-BP neural network, which takes into account the effects of the day-ahead load demand, real-time electricity price and other factors on the user’s electricity load. Based on the set real-time tariff, the electricity consumption plan of the load is adjusted, and finally, test system is developed. The final results indicate that the methodology proposed in this paper is beneficial to motivate users to participate in load demand response, and each user can make adjustments to the electricity plan according to their own actual demand for electricity.