In order to enhance the efficiency of flight operations and mitigate the negative impact of delays. The core causes of flight delays were analyzed in depth, covering the uncertainty of weather information, the fluctuation of flight status, and the congestion phenomenon during the departure time. Subsequently, a departure flight scheduling optimization algorithm was designed by integrating the inverse learning strategy, chaotic local search strategy and cross-variance strategy in the nutcracker algorithm. At the same time, a departure flight scheduling optimization model was constructed with the objective of minimizing the total delay cost, which provided a scientific basis and guidance for flight scheduling. Through the experimental validation of the optimal sequencing of departing flights within 18 hours in the real scenario of Xianyang airport, the results show that compared with the particle swarm optimization algorithm, the double-crossover genetic algorithm, the ant colony optimization algorithm, and the nutcracker algorithm, a reduction of at least 80% of the delay cost is achieved in the worst case scenario, and a reduction of at least 70% of the average delay cost, and achieves a significant reduction of at least 90% in variance. The result shows that our method has significant advantages in improving flight scheduling efficiency and reducing delays.