The generalized disjunctive programming (GDP) model has been proposed and applied in the past decade as an alternative to the mixed integer nonlinear programming (MINLP) model because it has the advantages of being straightforward when used in conditional modeling and being able to reduce the complexity in the sub-NLP. In this paper, we introduced an improved variant of the traditional GDP model, otherwise known as the object-oriented disjunctive programming (ODP) model. Such a method helps generate well-posed sub-NLP, thereby improving the solving process. A nested method combining the heuristic algorithm and gradient-based optimizer is also proposed to solve the GDP and ODP. It is a two-layer method, wherein a heuristic algorithm performs master iterations in the outer-loop when dealing with the Boolean variables, and a gradient-based NLP solver is applied in the inner-loop when dealing with the sub-NLP. Excellent performance has been demonstrated by applying the modeling and solving methods into the process synthesis of heat exchanger networks (HENs) and water networks (WNs).