Multidisciplinary design optimizations have shown great benefits for aerospace applications in the past. Especially in the last decades with the advent of high speed computing. Still computational time limits the desire for models with high level of fidelity cannot be always fulfilled. As a conse- quence, fidelity is often sacrificed in order to keep the computing time of the optimization within limits. There is always a compromise required to select proper tools for an optimization problem. In this final thesis work, the differences between existing weight modeling techniques are investi- gated. Secondly, the results of using different weight modeling techniques in multidisciplinary design optimization of aircraft wings is compared. The aircraft maximum take-off weight was selected as the objective function. The wing configuration of a generic turboprop and turbofan passenger aircraft were considered for these optimizations. This should aid future studies of wing shapes in early design stages to select a proper weight prediction technique for a given case. A quasi-three- dimensional aerodynamic solver was developed to calculate the wing aerodynamic characteristics. Various statistical prediction methods (low level of fidelity) and a quasi-analytical method (medium level of fidelity) are used to estimate the structural wing weight. Furthermore, the optimal wing shape was found using a local optimization algorithm and is compared to the results found using a novel optimization algorithm to find the global optimum. The quasi-three-dimensional aerodynamic solver was validated using experimental data and other available aerodynamic tools. Compared to the results generated by other tools, the developed solver has a wider range of validity. Most important of all, it is up to 10 times faster and the results show good agreement with other data. Several test cases were used to prove the robustness and effectiveness of the global optimization algorithm. A comparison of the different weight estimation methods indicated that the lower level fidelity methods are insensitive for some wing parameters. The results of the optimizations showed that the optimum wing shape is affected by the used weight modeling technique. Use of different weight prediction methods strongly affects the computational times and the convergence history. The global optimization algorithm was able to find the global solution for the wing shape optimization. However, the search for the global optimum comes at a cost: the computational time is significantly larger.