Along with other nitrogen oxides, nitric oxide (NO) is a regulated air pollutant that is primarily produced as a result of combustion processes. This pollutant is produced in every combustion system that uses oxidizer mixtures containing N-2 and/or fuels that contain organically bound nitrogen. To design combustors to minimize NO emissions, high-fidelity computational models for NO production and NO/NO2 interconversion are required. To improve model performance and understanding of NO production pathways, a computational parametric study is performed to investigate the effects of fuel chemistry, reaction temperature history, and inert gas dilution on NO production in methane and ethylene combustion. Predictions using popular models from the literature are compared for premixed laminar flame conditions as well as against previously reported stirred reactor measurements. Differences in the predictions and their relationships to the NOx submodel and hydrocarbon chemistry are investigated. We find that significant differences in rich flame conditions occur as a result of differences in the CH kinetic pathways and parameters within the hydrocarbon models, while NO production in lean and stoichiometric flames is sensitive to differences in the assumed Zeldovich (thermal) NO submodel parameters. As a result of uncertainties in the spectrum of time scales and the degree of mixing in stirred reactors, jet-stirred reactor data do not provide sufficient constraints on results to resolve the existing model differences. Well-defined time history evolution of NOx formation data are shown to be critical for developing improved characterization of the relative importance of the different submechanisms on NOx formation at gas turbine conditions.