This paper develops a cloud native thermal processing simulation platform to address fragmentation challenges in traditional tools, such as disconnected tool chains and limited cross-disciplinary collaboration. The platform implements a three-tier architecture comprising an interactive Process APP layer for task management and visualization, a Solver layer handling multiphysics computations and microservices, and a Resource layer orchestrating Kubernetes-managed GPU/CPU infrastructure. Innovations include a scientific workflow engine that enables drag-and-drop simulation customization, supported by a plugin-based framework that integrates commercial and open source solvers through REST/gRPC interfaces. Unified data governance using YAML and HDF5 standards ensures consistency across simulation modules, while cloud-native resource management provides elastic scaling capabilities. Validation confirms the platform’s ability to streamline parameter configuration via modular interfaces, facilitate multi-solver coupling for scenarios like casting solidification-stress analysis, and enable containerized distributed deployment. Future research will focus on deployment optimization and systematic methodology development for the thermal processing domain.