Abstract: In order to obtain a practical and high-precision forecasting method of spot market clearing price for electricity, we attempted to apply random forest regression (RFR) to the prediction of clearing price in spot market in this paper. Firstly, the feature importance analysis function of RFR was used to filter the input feature of historical clearing price and power load. Then we established a prediction model of market clearing price based on RFR, and the parameters of model were determined by grid search and cross validation. Finally, the experiment was carried out to compare this model with methods based on classification and regression tree, support vector machine regression and artificial neural network on the basis of the public data in Nordic spot market. The results show that the average prediction accuracy of the proposed prediction method is improved by at least 25% compared with other methods, and the prediction effect is relatively stable. Meanwhile, the application of input feature selection method can further improve the prediction accuracy.