An optimization method for sound absorption of gradient(multi-layered) sintered metal fiber felts is presented. The theoretical model based on dynamic flow resistivity is selected to calculate the sound absorption coefficient of the sintered metal fiber felts since it only requires three key morphological parameters: fiber diameter, porosity and layer thickness. The model predictions agree well with experimental measurements. Objective functions and constraint conditions are then set up to optimize separately the distribution of porosity, fiber diameter, and simultaneous porosity and fiber diameter in the metal fiber. The optimization problem for either a sole frequency or a pre-specified frequency range is solved using a genetic algorithm method. Acoustic performance comparison between optimized and non-optimized metal fibers is presented to confirm the effectiveness of the optimization method. Gradient sintered metal fiber felts hold great potential for noise control applications particularly when stringent restriction is placed on the total volume and/or weight of sound absorbing material allowed to use. An optimization method for sound absorption of gradient (multi-layered) sintered metal fiber felts is presented. : fiber diameter, porosity and layer thickness. The model predictions agree well with experimental measurements. optimization problem either either a sole frequency or a pre-specified frequency range is solved using a genetic algorithm method. Acoustic performance comparison between optimized and non-optimized metal fibers is presented to confirm the effectiveness of the optimization method. Gradient sintered metal fiber felts hold great potential for noise control applications par ticularly when stringent restriction is placed on the total volume and / or weight of sound absorbing material allowed to use.