Composite materials are being used in many fields of industry day by day. With this increasing interest in composites, the methods of joining composites have also become the focus of attention. Mechanical fasteners cause damage to the composite, increase in weight, and stress accumulation in the joint area. Recently, joining composites with adhesives has attracted the attention of researchers. In this study, glass fiber reinforced polymer composites were combined with two paste adhesive thicknesses, using two brands of paste adhesives as fast-curing and slowcuring, and three different peel plies and the effects of these three different parameters on the bonding strength were investigated both experimental and numerical analyses. In the experimental part of the thesis, glass fiber reinforced polymer composites were produced by the vacuum infusion method. The surfaces modified with different peel plies were combined with two different paste adhesives. A single-impact shear test was performed to examine the bond strength. As a result, it was observed that the fast-curing paste adhesive showed better performance in bond strength. At the same time, it has been experimentally demonstrated that the paste adhesive thickness of 0.6 mm has a positive effect compared to the paste adhesive with a thickness of 0.4 mm. It has been observed that the different peel plies used did not make a critical difference in the bond strength. In the numerical part of the thesis, six different regression models were used to model the shear strength of adhesive-bonded composites and then an optimization study was carried out by selecting the two best regression models that accurately express the physical model. Using the stochastic optimization methods, Differential Evolution and Nelder Mead algorithms, the optimum shear strength values possible with the existing parameters were found. This thesis contributes to determination of the bonded samples with the highest shear stress value by determining the optimum parameters