Halo is one of the most important basic elements in cosmology simulation, which merges from small clumps to ever larger objects. The processes of halos’ birth and merging play a fundamental role in studying the evolution of large-scale cosmological structure. In this paper, a visual analysis system is developed to interactively identify and explore the evolution histories of thousands of halos. In this system, an intelligent structure-aware selection method in What You See Is What You Get manner is designed to efficiently define user’s interesting region in 3D space with 2D hand-drawn lasso input. Then the exact information of halos within this 3D region is identified by data mining in the merger tree files. To avoid visual clutter, all the halos are projected in 2D space with MDS method. Through the linked view of 3D view and 2D graph, users can interactively explore these halos, including the tracing path and the evolution history tree.