This paper aims at modeling and optimizing a Middle East-based commercial natural gas liquid (NGL) recovery and fractionation plant, using a predictive process simulator. NGL units are known to be highly energy-intensive as steam-based heating and refrigeration-based cryogenic cooling are critical requirements for their operation. Indeed, these units govern the degree of profitability of gas plants especially during low natural gas price scenarios. As a result, this study explores the ways of improving the performance of NGL units through a deterministic optimization analysis. A steady state model of the plant is built using gPROMS process builder followed by validation using plant data to ensure the model accuracy. A mixed integer nonlinear programming optimization problem is formulated with the objective of maximizing the net revenue of the plants by means of manipulating various decision variables such as feed gas temperature, column operating pressure, feed stage location, reflux and boil up ratios subject to specific process constraints. Optimization problem is solved using outer approximation equality relaxation augmented penalty algorithm. It is determined that the process optimization yields an additional revenue of 4.1 MM USD annually due to ~22% increase in Liquefied Petroleum Gas (LPG) production, ~6% increase in Naphtha production, and ~16% reduction in steam consumption in the reboiler of the columns.