In order to improve the supply and demand of power market and realize the purpose of peak shaving and valley filling, the research on real-time pricing based on demand side management is deepening. In view of this, the main optimization methods and the latest progress were reviewed. Firstly, the social welfare maximization models, which aim to maximize the sum of utilities of all customers minus the power cost, in other words, to maximize the total social welfare, were summarized. Then, the methods of solving social welfare maximization models were discussed, which included the dual optimization methods, the alternating direction multiplier methods and the nonsmooth equations methods for KKT system. These methods were used for solving both the decision variables, which were electricity consumptions, and the Lagrange multiplier, which was the shadow price (as electricity price) of the optimization model. Next, the bi-level optimization model for the real-time pricing was introduced, where the power supplier was the leader and the power users were the followers. The basic properties and solution methods of this model were analyzed. Finally, several recent research problems of the optimization method for the real-time electricity pricing were discussed.