In order to improve the efficiency of selective disassembly sequence planning for repair, a selective disassembly sequence planning approach based on ant colony algorithm is proposed. Hybrid Precedence Graph(HPG) is constructed to describe the geometric constraints among constituting components and components' stability during disassembly. Then in order to find the target component accurately and quickly, disassembly set for the target is obtained based on HPG. Dynamical candidates set strategy based on precedence relations is adopted to constrain the searching space, and a selective Disassembly Sequence Planning(DSP) approach based on HPG is proposed. Finally, the validity and feasibility of the proposed approach is verified by a study case. Taking geometric constraint and components' stability into count, this approach could be more practical.