Benders decomposition algorithm to improve the computation for the decision problems involved in chemical processes is addressed. Using nonlinear duality algorithm. a large scale problem is decomposed into two stage problems, SUB and MASTER, respectively. The target problems are short-term scheduling and design problems, and long-term planning problems including capacity expansions. In the first issue, the net present cost involving design cost and operation cost with consideration of inventory is minimized. The latter addresses robust models that are insensitive to future uncertainty. The problems presented in this paper are so computationally complex that enhanced algorithms have been required. The effectiveness of the improved algorithm is illustrated through examples relevant to scheduling and design problem, and long-term capacity expansion problem, respectively.