China's road transportation industry accounts for a high proportion of carbon emissions, to a large extent, carbon emissions in road transportation are primarily caused by vehicle emissions. Against the background of China's carbon peaking and carbon neutrality strategies, it is particularly important for carbon dioxide emissions from road vehicles to be reduced. In order to predict the carbon emissions of light-duty vehicles during highway driving, the portable emission testing system (PEMS) was utilized to collect the carbon dioxide emissions, operating t conditions and environmental information of the test vehicle during driving on different grades of highway, and the optimal regression equations between vehicle carbon dioxide emissions and influencing factors were obtained, using speed, acceleration, gradient and air density as independent variables, and carbon dioxide emissions as dependent variables for correlation analysis and stepwise regression analysis. According to cluster analysis based on vehicle specific power (VSP) and operating speed, correlation analysis and stepwise regression analysis were conducted with speed, acceleration, slope, and air density as independent variables and carbon dioxide emissions as the dependent variable. The optimal regression equation between carbon dioxide emissions of test vehicles and influencing factors was obtained, and the influence weights were analyzed. The results show these as follows. The CO₂ model based on speed intervals of different levels of highways is more accurate than the model based on comprehensive speed intervals. Within the speed interval [20,40), CO₂ emissions are most significantly affected by the gradient, nevertheless, in other speed intervals, CO₂ emissions are most significantly affected by acceleration. Within the speed interval [2.0,20), speed is identified as the second most influential factor, and acceleration is assumed as the second significant factor in the speed interval[20, 40). However, within different speed intervals, carbon dioxide emissions are minimally influenced by speed and air density.