Future transition from traditional air traffic control to trajectory-based air traffic management is now discussed. Accurate trajectory prediction is indispensable for trajectory-based management. We developed a regression model to predict the flight time of flights bound for Tokyo international airport departing from Fukuoka airport. Aircraft are vectored to make an efficient arrival sequence to the airport in Kanto South sector. The flight time predicted by our model is the time required from the entrance of Kanto South sector to the airport. It is influenced by the decision made by controllers on the order of arrival sequence to the airport. The main repressors of our model are the modified positions of the other aircraft bound for the same runway, which are called traffic density data. The width of ninety five percent interval of prediction errors is more than three minutes narrower than that of ninety five percent interval of the original data. This shows the effectiveness of the inclusion of traffic density data as repressors, but our model still has a large prediction error, and further refinement is necessary.