Abstract: In order to reduce the mechanical properties and the curing residual stress of the thick composite caused by the excessive temperature peak during the curing process, a multi-objective optimization model based on the multi-physics coupling characteristics was developed to reduce the maximum curing temperature peak and the curing time. Firstly, a three-dimensional model which incorporated three typical sub-models including thermo-chemical model, resin viscosity model and resin flow model was established to investigate the development of temperature and thickness of laminate during curing process. The results of numerical model were compared with experiment data in reference and good accordance was obtained. Then, a multi-objective optimization method was applied to optimize curing process parameters by using a radial basis function neural network model (RBF) as the surrogate model. It is shown that the curing temperature peak has a nonlinear relationship with the first and second dwell temperature, which is related to the nonlinear characteristics of the curing process. In order to reduce the temperature peak, it is necessary to increase the first dwell temperature and reduce the second dwell temperature. Meanwhile, the dwell time should also be adjusted to shorten the total curing time. Compared to standard cure profiles, the proposed optimization method can significantly reduce the curing time and temperature peak for thick composite laminates.