Under the context of alliances, airline products have complex attributes and high risk of cooperation compared to individual flights. Differences in product attributes lead to different market share. Airline alliance partner selection and collaborative value calculation are based on accurate forecasting of airline market share. Therefore, based on the QSI (Quality of Service Index) indicators of Delta Airline, this paper made full use of the alliance transportation data and added indicators such as route competition degree and alliance cooperation rela- tionship to establish market share forecasting for international alliance route products. When fitting the nonlinear relationship between the eight factors and QSI value, considering that the parameter method needs to determine the form in advance, it introduced BP neural network to fit the nonlinear relationship between market share and factors. In order to solve the problem that the BP neural network falls into the local minimum value, the particle swarm optimization algorithm was introduced to determine the initial weight of the neural network and finally the QSI model of the PSO-BP algorithm was established. The results show that the QSI model based on PSO-BP al- gorithm can better predict the market share of international routes and the absolute error is below 1%.