Sodium (Na) metal is the most promising alternative for lithium metal as an anode material for the next-generation energy storage systems due to its high theoretical capacity, low cost and high natural abundance. However, huge volume change and severe dendrite growth of Na metal anode during repeated electrochemical stripping/plating result in rapid electrode degradation, low Coulombic efficiency, and even the risk of explosion caused by short circuit of the batteries. Because of its very soft nature, Na metal is susceptible to deform and degrade during both battery assembly process and electrochemical cycling. The volume and morphology change of the “hostless” Na severely aggravates the growth of Na dendrites. Thus, it is essential to seek an ideal host to support Na metal so as to maintain the high integrity of the electrode. Here we present the employment of a chemically engineered 3D porous copper (Cu) matrix with a cylindrical core-shell skeleton structure as a highly stable host for metallic Na anode, enabling a high Na areal mass loading of 70 mg cm^-2, through a facile and cost-effective surface treatment on commercially available Cu foam in forming high performance composite anode for Na-based batteries. We find that the unique surface characteristics of the as-obtained matrix can not only facilitate uniform impregnation and confinement of Na within the matrix pores promoted by the chemical interaction between Na and the matrix, but also can divert the Na plating from the matrix skeleton towards the Na reservoirs within the pores upon cycling. Benefiting from our delicately surface-engineered matrix as a host for Na, we demonstrate an excellent Na anode cycling stability in carbonate electrolyte without any additives over 100 cycles (300 hours) at a current density up to 2 mA cm^-2 with a high cycling capacity up to 3 mAh cm^-2. A stable rate performance up to 3 mA cm^-2 is also demonstrated. A full cell made of Na₃V₂(PO₄)₃ as cathode and the as-prepared surface-treated Na composite as anode is further tested, showing superior cycling performance and high rate capability (at 5C) compared with that using bare Na metal as anode. In addition to performance improvement, we systematically investigate the interfacial interactions between treated/untreated Cu matrixes and metallic Na to reveal how the surface property of the matrix skeleton greatly influences the electrochemical stability of Na anode.