Time-Sensitive Networking (TSN) is appealing to Industrial Internet of Things (IIoT) due to its support for deterministic real-time communication based on Ethernet. The deterministic transmission of time-sensitive traffic is achieved through a time-triggered mechanism, which requires precise schedule synthesis. However, existing approaches mostly focus on the static schedule, which cannot guarantee rapid response to dynamic requirements due to long runtime. Moreover, the multicast scheduling problem required for distributed applications in IIoT has been ignored. In this paper, we formulate the joint routing and scheduling problem for multicast time-sensitive traffic based on Integer Linear Programming (ILP) and extend it to cluster-ILP (CILP) to accelerate schedule synthesis. Firstly, an improved topology pruning is introduced to reduce the scale of the scheduling problem. Then flows are divided into several groups and schedules are calculated for flows in different groups incrementally. The evaluation results indicate that our proposed algorithm can reduce the runtime by 86.7% on average while ensuring the performance of resource utilization, thus increasing the schedulability for dynamic applications in TSN.