Probabilistic nonlinear first ply failure loads of flat composite panels and nonlinear buckling loads of curved composite panels with cutouts are estimated to provide the more reliable main load carrying structure in the renewable energy industry and offshore structures. The response surface method approximates limit state surface to a second order polynomial form of random variables with the results of deterministic finite element analyses at given sampling design points. Furthermore, the iterative linear interpolation scheme is used to obtain a more accurate approximation of the limit state surface near the most probable failure point (MPFP). The advanced first order second moment method and the Monte Carlo method are performed on an approximated limit state surface to evaluate the probability of failure. Finally, the sensitivity of the reliability index with respect to transformed random variables is investigated to figure out the main random variables that have an effect on failures.