By analyzing the production process of packaging lines and aiming at the multiple-target problem of the Delta parallel robot sorting with dynamically changing positions under complex multi-constraint conditions,a me- thod to optimize the sorting trajectories of the parallel robot is proposed based on an improved genetic algorithm and the thought of beats.By analyzing the sorting strategy,the practical problem is transformed into a similar traveling salesman problem (TSP),and an improved genetic algorithm is put forward according to the particularity of the sorting process.Moreover,the thought of beats is introduced,and the sorting trajectories in every beat are opti- mized by transforming the constraint conditions into a genetic operator combining the fission and merge operations of chromosomes.The test results show that,as compared with the traditional methods,the proposed method can ef- fectively shorten the average sorting stroke,with an average efficiency increase of 14. 76%.