Small-scale wind energy systems are increasingly popular for decentralized, renewable energy generation in urban settings, offering advantages like space efficiency, reduced grid dependence, and lower costs. These micro wind turbines constitute a promising solution for enhancing energy independence and reducing environmental impact. However, they face significant challenges, particularly at low Reynolds numbers, where laminar-to-turbulent transitions occur, and viscous forces are critical to flow behavior and performance optimization. Additionally, the operational conditions of these turbines are highly uncertain, with variations in wind speed, angle of attack, and turbulence levels significantly affecting their performance. This study proposes a robust design methodology for micro wind turbine airfoils, incorporating uncertainty quantification into the optimization process. The NACA 0012 airfoil serves as a case study, demonstrating that even slight uncertainties can lead to an important reduction in the average aerodynamic efficiency at a low Reynolds number. The methodology employs the Non-Intrusive Polynomial Chaos method, combined with a Kriging metamodel, to identify an optimal airfoil geometry. The results show an 8.3% improvement in mean aerodynamic efficiency, primarily due to a smoother pressure gradient on the suction side of the optimized airfoil, which reduces flow separation and enhances the overall performance.