In view of the low accuracy of gasoline engine intake flow prediction, the feasibility of support vector regression (SVR) to the intake flow prediction was analyzed, and an intake flow prediction model based on SVR was put forward. Based on the structural advantages of SVR, grey correlation analysis (GRA) was used to extract the feature vector of model, genetic algorithm (GA) was used to optimize and identify the model parameters, and the generalization performance and prediction accuracy of model hence improved. The model was trained and predicted by using the simulation test data of gasoline engine transient conditions, and the regression prediction function of SVR model was realized by using MATLAB/LIBSVM toolbox. The results show that the error between the predicted value of SVR model and the test value is within 2% and the prediction of inlet flow under the transient condition is effectively realized; Compared with conventional RBF neural network prediction model and BP neural network prediction model, it has higher prediction accuracy and is suitable for precise control of air-fuel ratio of gasoline engine under transient conditions.