In order to eliminate the nonlinear aerodynamic behaviour of a symmetric airfoil within the low angles of attack at low Reynolds number， this paper establishes a low Reynolds number airfoil optimization design procedure based on Gaussian process regression model and NSGA-Ⅱ multi-objective genetic algorithm coupled with RANS numerical method. Then this paper carries out the airfoil optimization design with the average cruise factor in the range of small angle of attack of multiple states as the objective function. The results show that the optimized airfoil exhibits pronounced geometric bump and effectively alters the position and type of the laminar separation bubble. The aerodynamic characteristics of the airfoil are significantly improved， and the nonlinear phenomenon of the lift coefficient within the small angle of attack range at Re=4×10⁴ basically disappears. Further analysis reveals that the bump of the optimized airfoil enhances the convective influence of the wall， which enhances the resistance of the incoming flow to the upward pressure gradient， thus improving the airfoil aerodynamic characteristics.