The use of optimization techniques is necessary in order to explore the full potential of laminated composite structures. Unfortunately, the computational cost of the optimization process can be very high when numerical methods are used to carry out the structural analysis. This work addresses the use of surrogate models to reduce the computational cost to optimize composite structures. The PSO algorithm is used for optimization and a sequence of surrogate models, based on the use of Radial Basis Functions, is used to approximate the structural responses. The accuracy of the proposed approach is assessed using a set of laminate optimization problems and very good results were obtained.