In order to solve the problem of poor prediction accuracy of departure flight time by BP neural network, a prediction model of departure flight time was established based on Support Vector Mac (SVM). Firstly, quantifiable factors affecting the departure time of flight are analyzed, and a prediction model of departure time of flight is built based on correlation analysis. The results of taxi-out time prediction based on SVM and BP neural network are compared and analyzed. The conclusions are as follows: (1) the taxi-out time of departing flights is strongly correlated with the number of launching flights in the same period, the number of taking off flights in the same period, the number of flights in the same period and the average taxiing time in one hour. It is correlated with taxiing distance, number of turns and delay time but not significantly, and is not correlated with the time period at the departure time. (2) The trend of prediction results based on SVM and BP neural network is consistent, and the accuracy of seven-tuple prediction results is the best considering strong correlation and moderate correlation factors. After introducing irrelevant factors, the prediction accuracy of the model will decrease. (3) The accuracy of the SVM-based taxi-out time prediction model is significantly higher than that of the BP neural network prediction model, and the prediction accuracy within the taxi-out time error range ±5 min can reach 98%.