© 2003-2012 IEEE. This paper presents the use of metaheuristics to support the design of mechanical systems, as a proposal of robust design methodology. The Particle Swarm Optimization algorithm (PSO) approach in the drive shaft made of composite materials, in order to obtain the optimal design parameters to support the loads specific to which this element will be subject. The problem in this type of design, is to find the optimal thickness of the piece, i.e. the number of layers and their respective orientation that will form it, for which there are too many possible combinations when working with composite laminates, the proposed methodology looks for the best possible solution according to the parameters of the problem. The main contribution of this work is the use of metaheuristics to solve complex design problems and difficult to solve by traditional methods. The experimental results lead to the conclusion that this type of implementations is very useful to solve problems of numerical optimization in the design with composite materials for diverse engineering applications.