Molecular simulations have been widely employed in the discovery of nanoporous materials, such as metal-organic frameworks (MOFs) and zeolite, for energy- and environment-related applications. To achieve simulation predictions with better accuracy, we herein present a collection of molecular models, including carbon monoxide (CO), carbon dioxide (CO2), carbonyl sulfide (COS), hydrogen sulfide (H2S), nitrogen (N2), nitrous oxide (N2O), and sulfur dioxide (SO2). These models, denoted as electrostatic potential optimized molecular models (ESP-MMs), are systematically developed to not only reproduce experimental vapor-liquid equilibrium but also have accurate electrostatic potential representation surrounding the molecules. Our results show that, with accurate electrostatic potential representations, ESP-MMs can offer improved predictions in a variety of adsorption properties for porous materials, including MOFs with open-metal sites and all-silica zeolites. Specifically, by using ESP-MMs, the binding geometry and adsorption energy landscape can be well captured. This enables these models to be employed to unravel the fundamental mechanism of gaseous adsorption in materials of interest as well as to facilitate the parametrization of adsorbent-adsorbate interactions. We also demonstrate that, combined with generic force fields for adsorbents, ESP-MMs can offer reasonable predictions in adsorption isotherms. Although these ESP-MMs use a relatively simple and nonpolarizable potential form for the sake of efficiency and applicability, their accuracy has been extensively validated in this study. Furthermore, the set of Lennard-Jones potentials with static point charges adopted for ESP-MMs can be readily implemented in all available simulation packages. We anticipate that these ESP-MMs can largely facilitate future computational studies of porous materials for gas separation and removal.