The clearing price in the electricity spot market is an important reference for market players to make bidding decision and evaluate revenue. The clearing price is affected by many factors and fluctuates frequently, which brings difficulties to the selection and evaluation of forecasting model. Firstly, the requirements of evaluation indexes for price prediction errors are summarized and the deficiencies of traditional indexes are analyzed. Then, in view of the different economic consequences caused by price prediction errors of different market players, a risk-based error assessment index is proposed. For operators and power generators, the error values during different time periods are given as weights. For retail companies, this paper analyzes and predicts whether the positive and negative price difference between day-ahead and real-time will affect their bidding strategy. For end users, the conncetion between forecast error and cost of electricity is analyzed. The index can accurately reflect the risk cost caused by the prediction error, and can be used to choose the appropriate prediction model according to the demand. Finally, the use of Guangdong spot market full-month trial operation data shows that the proposed method enables different players to accurately select the price prediction model that makes their decision-making risk the lowest. After the method is applied to industrial users’ decision-making, the cost of electricity purchase can be reduced by 1.76%~2.97%.