Multi-blade centrifugal fans find wide application across various fields, with their internal airflow exhibiting complex turbulent behavior in three dimensions. Historically, blade optimization relied on constant thickness airfoils, limiting the effectiveness of optimization efforts. However, marine organisms have developed airfoil structures with highly efficient drag reduction, offering a novel approach to optimizing multi-blade centrifugal fans. This study proposes an airfoil optimization design method utilizing variable-thickness airfoils to maintain consistent pressure surface leading-edge parameters. By integrating the Non-dominated Sorting Genetic Algorithm II with biomimetic optimization design, performance improvements are achieved. The study constructs an asymmetric bionic blade using a Bezier curve to fit the mean camber line of the blade. Experimental testing validates the optimized fan's performance, demonstrating the effectiveness of the proposed design approach in reducing the unsteady interaction between the impeller and the volute tongue. This reduction significantly diminishes sound pressure fluctuations on the blade surface. Notably, at the maximum volume flow rate, the optimized fan featuring the asymmetric bionic blade exhibits a remarkable enhancement, with a 10.5% increase in volume flow rate and a notable 1.7 dB reduction in noise compared to the original fan configuration.