In e-Science environments, the support for scientific workflows emerges as a key service for managing experiment data and activities, for prototyping computing systems and for orchestrating the runtime system behaviour. Supporting domain specific applications via a common e-Science infrastructure enables knowledge sharing among different applications, and thus can broaden the range of the application and multiply the impact of scientific research. However, most of the existing workflow management systems are driven by the domain specific applications; the applicability to different domains is limited. In this paper, we discuss possible solutions to this problem and present our research in an ongoing project: Virtual Laboratory for e-Science (VL-e). Agent technologies are used to encapsulate the intelligence for problem solving strategies and for workflow orchestration.