Under the mobile production mode of flow production line, the establishment of pulsating production line in general assembly workshop has become a development trend, and it is particularly important to solve the balance problem of pulsating assembly line in workshop. Considering the constraints such as assembly line beat and assembly sequence, aiming at minimizing the assembly time and the number of human resources, this paper constructs a multi-objective and multi-constraint balance problem model of final assembly pulsating production line, and designs a non dominated sorting genetic algorithm to solve it. Combined with the actual data of a workshop, the program is developed, and the final results are analyzed in detail. It shows that the algorithm can greatly improve the assembly efficiency and personnel utilization, and verify the feasibility and effectiveness of the algorithm.