Test or simulation data in engineering applications are usually expensive to collect. As a result, exhaustive search for optimal solutions in the entire design space is hardly practical in most engineering optimization problems. Recent advances in Bayesian optimization have shown promising results in solving engineering problems and require only small DOE data sets. Based on the existing methods developed in the literature, we propose an new intelligent sampling framework that aims to solve larger-scale optimization problems, ie, engineering applications with multiple objectives, high dimensional design space, and nonlinear/non-convex constraints. The capability of the proposed framework is demonstrated in three numerical examples.