The numerical prediction of production from stimulated Ultra-Low Permeability (ULP) media is highly correlated with the type of fracture model used in the simulator. Although there are some general rules about the applicability of these fracture models, there are no guidelines for the a-priori selection of an appropriate model -- generally, the fracture model is selected based on the preference and/or familiarity of the person performing the modeling, rather than some "more technical" criteria. In that light, this work presents an effort to provide guidance for fracture model suitability for cases from the Eagle Ford, Bakken, Three Forks, and Wolfcamp formations. In this work, production data from multiple wells in the aforementioned reservoirs are history-matched using models commonly available in commercial reservoir simulators. We evaluate the ability of the equivalent continuum model (ECM), the dual porosity, the dual permeability, and the multiple interactive continuum model (MINC) to represent these wells. We determine that a correlation exists between the choice of the fracture model and the reservoir. However, the results of the study do not provide a sufficiently strong indication of model superiority which would support authoritative guidelines about applicability for a particular reservoir. Ultimately the choice of the most representative model depends on the particular well. The proposed recommendationsprovide guidance onthe appropriate fracture model to represent the studied reservoirs, thus enhancing their usefulness in the completion design process and the evaluation and prediction of production.