The invention relates to a short-term commercial power load prediction method based on SARIMA-GRNN-SVM. The short-term commercial power load prediction method comprises the following steps: analyzinga load curve to obtain influence factors of commercial power load fluctuation; constructing a single prediction model for commercial power load time series prediction and multi-factor regression prediction; constructing an SVM (Support Vector Machine) model, performing parameter optimization and training on the SVM model by utilizing the training sample, and performing parameter optimization on the SVM model through a grid search and cross validation method; and inputting prediction values of the prediction day obtained by the SARIMA model and the GRNN model into the trained SVM model to obtain a commercial power load prediction value of the prediction day. According to the method, the problem that a prediction result is likely to have a large error due to the fact that a single predictionmodel cannot comprehensively consider periodic changes and influence factors of commercial loads is solved, and the prediction accuracy and robustness are improved.