This work is to summarize the research results of feature-driven topology optimization methodology[6-14]developed at the Engineering Simulation and Aerospace Computing (ESAC) Laboratory of Northwestern Polytechnical University,China in recent five years.The aim is to establish an efficient optimization tool that is CAD-oriented and more adapted to mechanical designs than the density-based standard topology optimization method.Challenging problems are thoroughly addressed about definitions of design features,design domain and design variables;feature operations and topology evolutions;effects of design-dependent loads;elimination of grey material presence;optimization of Dirichlet boundary conditions and the management of computing grid.Comparative studies are made with the existing density-based method and discrete level-set method to highlight the performances and capabilities of the proposed methodology through abundant examples.