The　distribution　of　grains　plays　a　crucial　role　in　determining　the　strength　of　polycrystalline　copper　when　the　grain　size　is　constant.　Herein,　the　mechanical　properties　of　homogeneous　nano-grained　(HNG)　and　gradient　nano-grained　(GNG)　coppers　with　different　grain　distributions　are　examined　using　molecular　dynamics　(MD)　simulations.　The　HNG-ordered　structure　has　all　triple　junctions　(TJs),　whereas　the　random　structure　contains　many　quadruple　and　quintuple　junctions.　When　grain　size　is　below　the　critical　size　in　the　inverse　Hall-Petch　relationship,　the　high-density　TJs　in　the　HNG-ordered　structure　effectively　inhibit　grain　boundary　(GB)　softening　compared　with　the　random　structure,　leading　to　higher　strength.　However,　when　grain　size　is　above　the　critical　size,　subgrains　are　produced　inside　the　large　grains　due　to　dislocation　slip.　In　addition,　disordered　atoms　in　HNG-random　structure　are　stacked　in　the　quadruple　and　quintuple　junctions,　resulting　in　thicker　GBs.　This　triggers　grain　boundary　migration,　and　forms　more　subgrains　at　GBs.　Subsequently,　grain　boundary　sliding　and　grain　rotation　of　subgrains　induce　partial　recrystallization　in　the　structure.　This　consecutively　triggered　deformation　mechanism　leads　to　extra　strengthening　in　the　random　structure.　Further　research　indicates　that　combining　small-grained　ordered　and　large-grained　random　structures　can　be　a　new　approach　to　effectively　strengthen　GNG　materials.