There is a growing need for highly accurate in silico and in vitro predictive models to facilitate drug discovery and development. Results from in vitro permeation studies across the Caco-2 cell monolayer are commonly used for drug permeability screening in industry and are also accepted as a surrogate for human intestinal permeability measurements by the US FDA to support new drug applications. Countless studies carried out in this cell line with published permeability measurements have enabled the development of many in silico prediction models. We identify several common cases that illustrate how using Caco-2 permeability measurements in these in silico and in vitro predictive models will not correlate with human intestinal permeability and will further lead to inaccuracies in these models. We provide guidelines and recommendations for improving these models to more accurately predict clinically relevant information, thereby enhancing the drug discovery, development, and regulatory approval processes.