With the continuous increase in carbon emissions from road transportation and significant regional differences, understanding the spatio-temporal evolution patterns and influencing factors of carbon emissions from road transportation in China is crucial for energy conservation and carbon reduction in this sector. Based on the calculation of carbon emissions from road transportation in 31 provinces of China from 2003 to 2022, this study uses slope values, Theil index, and integrates XGBoost-SHAP to construct an interpretable machine learning model to analyze the temporal characteristics, spatial differences, and influencing factors of carbon emissions from road transportation in China and its eastern, central, and western regions. The results show that: The combined carbon emissions from small and medium-sized passenger cars and heavy-duty trucks account for over 80%, while the proportion of light-duty trucks first increases and then stabilizes, and that of medium-duty trucks continuously decreases. The proportions of medium and large-sized passenger cars, buses, and taxis are relatively low and show a downward trend. The carbon emissions from road transportation in China exhibit a phased characteristic of "rapid growth-decelerating growth". The eastern region had a relatively fast growth rate in the early stage, but it has been continuously decelerating in the later stage; the central region's growth rate surpassed that of the eastern region in the later stage; the western region has a low base but the fastest growth rate. The overall spatial pattern of carbon emissions from road transportation in China is "expansion in the east-increase in the central-stability in the west", with high-emission continuous belts formed successively in the eastern coastal areas and the central hinterland. The differences in carbon emissions from road transportation between regions are gradually decreasing, while the differences within regions are significantly increasing. The differences among provinces within the eastern and western regions are significant, while those in the central region are the smallest and remain stable. (4) At the national level, the turnover volume of road freight is the primary factor and has a significant driving effect on carbon emissions from road traffic, followed by the total population and the number of motor vehicles per 10,000people, moreover, each significant factor shows a synergistic effect of increasing emissions. At the regional level, the turnover volume of road freight remains the most significant factor in the eastern and western regions, but it is not significant in the central region. The contribution rate of motor vehicles per 10,000people and per capita GDP in the central region has significantly increased compared to the eastern and western regions.