Evaluation method of exhaust gas simulation capable of simply and appropriately evaluating the validity of the simulation is provided. In analysis data, an analysis amplitude curve is calculated in which a change in the concentration of virtual exhaust gas at the observation point in the converged pipe portion is plotted, and an analysis time interval between the zero point and the reference point in the analysis amplitude curve is plotted. In actual measurement data, an actual amplitude curve is provided in which a change in the specific gas component at an observation point is measured with time, and an actual time interval is provided in which a time interval from a zero point to a reference point in the actual amplitude curve. The analysis data is determined as valid when a difference between the analysis time interval and the actual time interval is within a predetermined correlation range.