Moderate or intense low-oxygen dilution (MILD) combustion is a promising technology for simultaneously reducing NO emission and improving thermal efficiency. To apply MILD combustion in industrial production, computational fluid dynamics (CFD) modeling plays an efficient role in design and optimization. To reduce the computational time while maintaining the predictive accuracy, a valid global reaction mechanism (GRM) is necessary. In this paper, an accurate and robust GRM was proposed for methane under MILD combustion conditions, especially without the condition of highly preheated air. The adequacy of the proposed GRM was first compared with experimental data from a bench-scale MILD combustion furnace, where ambient-temperature air was used. Subsequently, experimental data from an industrial-scale MILD combustion furnace, where air was slightly preheated to 130 degrees C, was employed to further validate its capability for practical application. As compared to the previous GRMs, the present proposed GRM exhibits improved predictive accuracy in terms of flame temperature, oxygen concentration, and carbon monoxide concentration, for both bench-scale and industrial-scale MILD combustion cases. With growing interest being focused on MILD combustion using low temperature or even ambient-temperature air, the present proposed GRM is expected to be adopted by CFD users to design and optimize MILD combustion processes with reliable results and less computational time.