At present, the high-performance cluster system has been widely applied to multitask and multiuser data processing procedures. However, computation loads can be influenced by the job scheduling optimization strategy (JSOS) of the cluster system, which can cause imbalance during job scheduling process, job starvation, and resource fragmentation. This situation can further result to problems, such as dissatisfactory resource utilization and lengthy job response, turnover, and completion times. First, a double hierarchical job scheduling model was proposed in this study to optimize the job scheduling strategy of the cluster system. Second, this study analyzed the hierarchical tasks in the scheduling model and the factors, namely, resource utilization and job completion time that influence them. The reasonability of the JSOS was also verified. Finally, the optimization strategy for job scheduling was compared with the first-come, firstserved (FCFS) and FirstFit strategies. Result shows that compared with FCFS and FirstFit strategies, the proposed job scheduling strategy increased resource utilization by 6.3% and 0.8%, and reduced average response time by 22.05% and 1.12%, average turnover time by 9.82% and 2.01%, and completion time by 10.45% and 1.11%, respectively. Thus, the proposed double hierarchical job scheduling strategy not only improves system resource utilization but also reduces job response, turnover and job completion times. The experimental result is consistent with the expected requirements, and the study provides a feasible scheme for the job scheduling optimization problem in the cluster system.