An improved methodology, with considering experimental results and numerical models, is proposed to find globally optimum designs of composite laminates subject to given in-plane loads for a maximum fatigue life. For this purpose, the Fawaz-Ellyin parametric fatigue-life prediction model was coupled with a global optimization technique, the particle swarm optimization, which is a metaheuristic optimization algorithm. The number of distinct fiber orientation angles and the number of plies in each lamina were predetermined. The fiber orientation angles in each layer were taken as design variables and determined in the design process. A number of problems were solved to demonstrate the efficiency and reliability of the method proposed.