This dissertation investigates the use of Genetic Programming for the structural optimization of carbon clusters. The main objective concerning computations that describe the interactions of a cluster is to find the arrangements of atoms corresponding to the lowest energy, since these are the most likely candidates to be formed. It has been recently introduced in the area of Computational Intelligence a new optimization method, called Optimization by Genetic Programming (OGP), showing promising results for several benchmark cases. Based on these results, the present work aimed at the application of OGP for the geometry optimization of carbon clusters. For comparison purposes, independent optimizations using the standard genetic algorithm (GA) approach were carried out. Several optimization trials were performed using both GA and OGP in order to find the best geometries of carbon clusters with size ranging from 5 to 25 atoms. The energy was calculated using the Morse potential. Resulting energies and geometries were compared to previously published results. Both GA and OGP were able to find the global minimum for the smaller clusters. However, upon increasing the number of atoms, the OGP presented better results compared to the GA. Concerning the execution time for each evaluation, the GA is significantly faster than the OGP due to its direct representation of the positions of atoms of a cluster in a chromosome. However, the superiority of the OGP results compared to the GA results suggests that an effort towards the improvement of the implementation of OGP could lead to a very powerful optimization tool to be used by the scientific community.