In order to reduce flight delays caused by various reasons and improve the robustness of flight plans, based on the analysis of the flight delay data of a certain airline in winter and spring flights for a certain month and ten days, an optimization model of flight plan based on the improvement of robustness is established. The model considers robustness factors when formulating flight plans, and controls the increase in flight operating costs by constraining costs. In order to verify the accuracy of the model, this paper uses Lingo to perform modeling and brings in ten-day flight data for verification. The experimental results show that the robustness is 0 when the flight cost increase rate is 0, and the robustness has the maximum solution when the flight cost increase rate is 0.2. The results are normal on the eighth day of ten days, and only two days are due to missing data. Smaller solution, preliminary verification of the model has been completed.