Numerous industrial applications in mechanical, petroleum, chemical and nuclear arena require the interaction of two-phase flow in pipelines to be predicted for efficient and economic sizing and operation of equipment. While multiple empirical models have been developed for predicting the pressure drop in two-phase gas-liquid flow, they usually generalize the solutions and can occasionally be erroneous. Computational Fluid Dynamics (CFD) simulations can estimate pressure drops more accurately by numerically solving continuity and momentum equations on a case by case basis. These simulations provide a virtual laboratory experience where fluids and their interfaces can be analyzed spatially and dynamically. While several CFD studies on two-phase flow have been conducted in the past, they are usually limited to specific flow patterns or Reynolds number ranges. Either way, replicating them requires multiple trials incorporating different models and extensive computation time, with limited checkpoints to validate the findings. Thus, there is still a need for extensive work of two-phase flow CFD models which consider a balance between computation time and accuracy. The purpose of this study is to simulate two-phase gas-liquid flow using CFD models for 2D and 3D in smooth and rough pipes. Frictional pressure drop is calculated for dispersed, annular, stratified and slug flow patterns. In this study, these CFD simulations have been validated experimentally for pipe diameters between 0.027m to 0.05m. Mesh independence and multi-dimension analysis have been studies to balance computation time and accuracy, and hence provide an optimized CFD model to predict pressure drop for two-phase gas liquid flow.