Small wind turbines (SWTs) are a prominent renewable energy technology for decentralized power generation. Blade material and its profile are vital parameters for the aerodynamic performance of SWTs. Traditionally E-glass fibre-reinforced composites (FRCs) are the widely accepted material for developing SWT blades. However, its application is limited by moderate tensile and fatigue properties. Alternatively, other FRC materials such as carbon, basalt and natural fibre composites are proposed as future materials for SWT blades. However, individual materials are observed to satisfy the requirements partially. Therefore, the hybridization of these materials, particularly Glass/Carbon composites is foreseen as a prospective solution for developing cost-competitive and high-strength SWT blades. There are various studies performed to obtain optimized glass/carbon hybrid composites. However, overall material properties required for SWT blades such as low cost, lightweight, moderate flexural strength and higher tensile and fatigue strengths have not been considered simultaneously during the optimization process. This work presents multi-objective optimization of Glass/Carbon hybrid composites using extreme mixture design response surface methodology (RSM) for small wind turbine (SWT) applications. The weight percentages of glass and carbon fibres are optimized to achieve desired material properties for SWT blades. The experiments are planned using extreme mixture design RSM and the regression models for desired material properties are developed with a 95% confidence level. RSM-based desirability function is employed to perform multi-objective optimization. Maximum composite desirability of 93.5% is achieved with optimal proportions of 37.9% and 27.1% for glass and carbon fibres respectively. An adequate tensile, flexural and fatigue strengths of 486.02, 435.41 and 316.27 MPa respectively are obtained for optimized glass/carbon hybrid composite at an optimum cost of 2228.76 Rs/Kg and density of 3.39 g/cm3. The regression models and optimization results are validated through a confirmation experiment with an error of less than 6.1%.