Vertical axis wind turbine (VAWT) has a rotating axis perpendicular to the wind direction. This type of wind turbine that is suitable for urban environments has low wind direction dependency and noise. In this research, a novel surrogated approach for optimizing a VAWT is proposed, used, tested, and verified, which is not reported in literature. The proposed method consisted of 3D computational fluid dynamics (CFD) analysis of wind flow through the wind turbine with FLUENT software by solving the unsteady turbulent equations. However, 3D CFD analysis was time and cost consuming to obtain the output result (power coefficient) from input values (airfoil chord length, pitch angle, and tip speed ratio as turbine design variables). Thus, artificial neural network (ANN) was applied to obtain weight functions to correlate FLUENT software inputs and outputs after learning process. Finally, genetic algorithm was used for maximizing the turbine power coefficient considering three defined design variables. The optimum value of power coefficient was improved to 0.244, and the optimum values of design variables for blade chord length, blade pitch angle, and blade tip speed ratio were 0.218, −0.453, and 1.24, respectively. This novel surrogated method reduced the computational time and cost of VAWT optimizing considerably.