Abstract: For the typical discrete manufacturing characteristics of aerospace products with multiple varieties and variable/small batch sizes, a robust scheduling model is established with optimization objectives of minimizing deviation time at the start of a process, order delivery deviation time, robustness indices, and maximum completion time efficiency, with consideration of production disturbance such as emergency order insertion, process change and process deviation. A robust evaluation model is developed to evaluate the robustness of scheduling schemes with production disturbance. A master-apprentice evolutionary algorithm based on the agent mechanism is designed to achieve multi-objective comprehensive optimization of production disturbance. To verify the performance of the robust scheduling method, performance comparisons are conducted between the original master-apprentice evolutionary algorithm and classical particle swarm algorithm under three uncertain processing time scenarios. Results show that the proposed algorithm significantly outperforms other intelligent optimization algorithms in scheduling performance under the condition of considering different types of uncertainties, effectively balancing the efficiency and robustness of the scheduling process.