Optimization of 10.6m diameter, constant speed and stall regulated horizontal axis wind turbine was performed for Gokceada location in Turkey using a genetic algorithm. Optimizations were performed by maximizing the annual energy production of the turbine which was calculated using a blade element momentum theory code. Predictions by the code were validated by comparing them with experimental data and computational fluid dynamics solutions obtained using different turbulence models. Chord length and twist angle distributions along the blade span, airfoil profiles for the root, primary and tip sections of the blades, and pitch angle of the blades were used as design parameters. A micro-genetic algorithm with a population size of five individuals was used for the optimizations. Two-and three-bladed turbines were considered for rated power of 20 and 30 kW. Optimizations performed for 5000 generations showed that two-bladed turbines could produce as much energy as three-bladed ones at the expense of rotating faster and experiencing higher blade loading. By having a lower solidity, however, two-bladed designs produced lower thrust forces.