Abstract: A large number of new market participants have emerged in the electricity spot market and have been granted certain pricing rights. This development significantly improves the flexibility, volatility, and complexity of the clearing price, thereby elevating the risk of abnormal prices. To ensure the safe and stable operation of the electricity spot market, it is necessary to develop an abnormal price warning method to mitigate potential abnormal price risks in advance. Firstly, the abnormal performance and sources of the clearing price are analyzed. A pre-assessment indicator system for the state of the clearing price evaluation is constructed from seven aspects: price level, price volatility, market structure, participant behavior, network operation, system management, and external factors. Secondly, based on the pre-assessment indicator system, two different warning methods are proposed: a BP neural network warning method with dimensionality reduction using the linear discriminant analysis method and a BP neural network warning method with dimensionality reduction using the local tangent space alignment method. These methods apply BP neural networks to assess the possibility of subsequent price abnormalities after dimensionality reduction of the data. Finally, the proposed abnormal price warning methods are applied to the actual operational data of an electricity spot market in a certain region. The results indicate that both warning methods are effective for abnormal price warnings, and the dimensionality reduction methods significantly improve computational efficiency within a given accuracy loss.