In order to increase the efficiency of measures targeting combustion noise in complete vehicle development, a new sound quality evaluation method was developed by combining two known methods. One of these is the time domain combustion noise separation method (T-CNSM). The procedure of this method is as follows. Firstly, simultaneous measurement of vehicle interior noise and in-cylinder pressure of each cylinder is carried out under acceleration. Afterwards, the contribution of combustion noise in the time domain is separated from the vehicle interior noise by the application of digital signal processing, using Fourier transform and multiple regression analysis, to the measured data. This method, the T-CNSM, has made it possible to remove other noise sources, focusing exclusively on combustion noise. Furthermore, the T-CNSM has enabled a sound quality check of combustion noise using a headphone-playback system. The other method is the objective sound quality evaluation method for combustion noise in vehicle interior. Subjective evaluation tests using Scheffe’s paired comparison method were conducted to fourteen NVH (noise, vibration, and harshness) development experts via a headphone-playback system. Multiple types of combustion noise in vehicle interiors obtained by the T-CNSM were used for the presentation sound. Based on the results of the subjective evaluation tests, the dissonant combustion noise index (DCNI) was developed as a new objective evaluative index for the sound quality of combustion noise using fluctuation intensity as a psychoacoustic metric expressing intensity of perceived fluctuation. The evaluation method combined the DCNI with the T-CNSM, enabling a highly precise sound quality quantification of combustion noise in vehicle interior. This paper provides an overview of this evaluation method’s procedure series and discusses the validity of this method when applied to the comparison of combustion noise sound quality in three small Japanese Kei car models.