To improve polishing quality and cope with the shortage of skilled workers for aluminum wheel-hub surface polishing, an automatic surface polishing system with hierarchical control based on the teaching-playback method was presented. Multi-axis cutter location data (CL data) were generated with the teaching method. First, a helical tool path and a flexible polishing tool were adopted to achieve high quality and high efficiency; next, the initial irregular data were processed into continuous polishing CL data. The important factor affecting polishing quality, namely the interpolation cycle in the multi-axis CL data was calculated based on a constant removal rate. Results from polishing experiments show that the quality of automatic machine polishing is better and stabler than manual polishing. To improve polishing quality and cope with the shortage of skilled workers for aluminum wheel-hub surface polishing, an automatic surface polishing system with hierarchical control based on the teaching-playback method was presented. Multi-axis cutter location data (CL data) were generated with the teaching method. First, a helical tool path and a flexible polishing tool were adopted to achieve high quality and high efficiency; next, the initial irregular data were processed into continuous polishing CL data. cycle in the multi-axis CL data was calculated based on a constant removal rate. Results from polishing experiments show that the quality of automatic machine polishing is better and stabler than manual polishing.