In order to target subsequent optimisation measures to reduce the negative impact of flight delays in multi-airport terminal areas and to improve the operational efficiency of each airport in a multi-airport system, a study on the prediction of flight delays in multi-airport terminal areas was conducted. Firstly, the impact of multi-airport terminal area traffic dynamics on flight delays was considered. Based on the analysis of multi-airport terminal area traffic dynamics, six indicators describing terminal area traffic dynamics were established. Then, a back propagation (BP) neural network flight delay prediction model were constructed, taking the terminal area traffic situation indicators, flight information and weather environment data as input and flight delay time as output, and the BP neural network was optimized using particle swarm optimization algorithm (PSO) for training. The results show that flight delay prediction based on multi-airport terminal area traffic situation can effectively improve the prediction accuracy, and the prediction accuracy of the prediction model by the particle swarm optimized BP neural network is higher than that of the general BP and genetic algorithm and back propagation (GA-BP) models considering traffic situation.