With the development of multiprocessor technology, multiple processors are increasingly used in embedded real-time systems. The parallelism of multiple processors makes the worst-case response time (WCRT) analysis of multiprocessor real-time systems complicated. In the widely used global fixed-priority (GFP) scheduling, some effective approximate WCRT analysis methods have been proposed. These methods mainly focus on optimizing overall procedure or carry-in workload and assume that the whole carry-out workload of a high-priority task cause interference on the analyzed job. However, through our careful observation of specific examples, the partial carry-out workload of a high-priority task and the analyzed job may be executed in parallel. In other words, the part of parallel carry-out workload of a high-priority task could not cause interference on the analyzed job. In this article, we propose an improved WCRT analysis method by optimizing carry-out interference estimation for GFP scheduling, resulting in a more accurate upper bound of WCRT than the recent advanced methods. Experimental results show that our proposed method is superior than the recent advanced methods in terms of the acceptance rate of the system.