In　order　to　prevent　robot　colliding　to　human,　and　improve　robot　safety　during　motion,　a　method　of　robot　safe　motion　planning　based　on　minimized　danger　index　is　proposed.　The　motion　planning　of　robot　is　divided　into　four　steps　which　include　global　path　safe　planner,　real-time　safe　trajectory　planner,　real-time　anti-collision　controller　and　goal　re-searching.　The　danger　degree　can　be　estimated　based　on　the　robot　inertia,　the　distance　and　relative　speed　between　robot　and　human　during　every　planning　step,　and　it　described　as　danger　index.　The　motion　of　robot　can　be　planned　by　minimizing　the　danger　index.　A　mathematic　model　including　a　3　degree　of　freedom　robot　and　human　is　established.　Every　motion　planning　step　is　simulated　in　Matlab.　To　further　verify　the　validity　of　the　planning　algorithm,　a　safe　motion　planning　experiment　platform　of　industrial　robot　is　established,　and　safe　avoiding　collision　is　realized　based　on　minimized　danger　index.　The　simulation　and　experiment　results　show　that　this　robot　safe　motion　planning　method　based　on　minimized　danger　index　is　capable　of　ensuring　human　safety　in　coexistence　environment　of　robot　and　human.　©2015　Journal　of　Mechanical　Engineering