In this thesis study, the aim is to propose a methodology on the optimum stacking sequence design of carbon/epoxy composite laminates under various cyclic loading conditions for maximum fatigue life. In this respect, first, fatigue life prediction models, Failure Tensor Polynomial in Fatigue (FTPF), Fawaz-Ellyin (FWE), Sims-Brogdon (SB) and Shokrieh-Taheri (ST) are selected to investigate their prediction capabilities in multidirectional laminates and optimization capabilities in laminate design for maximum fatigue life. An experimental correlation study is performed for different multidirectional composite materials to evaluate the prediction capability of the models by comparing to each other. The predictions of the models give accurate and close results for all the composites in many lay-up configurations. Then, the optimum designs for maximum fatigue life are obtained for glass/epoxy composite laminate from the literature using different powerful hybrid algorithms to determine the optimization capability of the models. The results of the optimization imply that FTPF and SB models produce more consistent fatigue-resistant designs than FWE and ST models. After obtaining reasonable theoretical derivations, the methodology for fatigue-resistant design is applied to carbon/epoxy composite laminates under proper cyclic loading conditions. For this, first, quasi-static and fatigue strength properties of the carbon/epoxy laminates are determined by experimental procedure. Then, many problems including different design cases are solved using the FTPF model and hybrid PSA-GPSA algorithm, and multidirectional laminate designs with maximum fatigue life are determined. The results show that fatigue strength of the composite laminates can be seriously increased by appropriate stacking sequence designs.