The determination of the optimal overbooking policy is the core issue of overbooking, an important constituent of airline revenue management. In view of this problem, based on the binomial distribution overbooking model, this paper takes into a synthetic consideration the influence of the number of No-show as well as the number of DB on the total revenue. After a numerical analysis of the changes of the parameters of multivariate function by using MATLAB, a conclusion has thus been drawn of the relationship between the optimal booking level of flights and the probability of passengers arriving at the airport, as well as the handling of the expenses of DB passengers, with some effective suggestions to be given. Finally, an empirical analysis has been made of the model, thus verifying the simplicity and feasibility of the proposed method, and can be used as a reference for airlines to set booking level.