In the study of flexible job-shop scheduling problem（ FJSP）

a large-scale FJSP data set based on the traditional benchmark instance is designed to simulate the flexible job-shop scheduling scenario of the actual largescale production workshop. Aiming at the large-scale FJSP dataset

a multi-objective flexible job-shop scheduling model with the optimization index of maximum completion time

maximum machine load and total machine load is proposed. In order to avoid the algorithm falling into local optimum

an improved multi-group NSGA-Ⅱ algorithm（ IMNSGA-Ⅱ） is proposed. The algorithm optimized the population initialization method and cross-strategy

and adopting different crossover methods for the genes of the machine parts in different populations. Effectiveness and feasibility of the algorithm are verified in the large-scale FJSP dataset.