With the continuous rise of carbon emissions from road traffic, the global greenhouse effect is constantly intensifying. In recent years, the dual model approach of factor decomposition models and carbon emission prediction models, combined in industrial and energy fields, has shown its advantages in revealing the key driving factors of carbon emissions and uncovering the path to carbon peak, but its application in the field of road traffic is still insufficient. A multi-scale emission inventory model is used to obtain road traffic carbon emissions data from 2001 to 2019 in China and the GDIM method is used to decompose the driving factors that affect carbon emissions (including GDP, road traffic energy consumption, unit energy carbon emissions, population total, road traffic per capita carbon emissions, per capita GDP, unit GDP energy consumption, and road traffic carbon emission intensity). Secondly, five-layered scenarios are designed to evaluate the emission reduction potential under different policy combinations; finally, the LEAP model is employed to simulate and predict the carbon peak situation of road traffic in China from 2021 to 2035. The results show that: (1) Among the driving factors, GDP was the most important factor affecting traffic carbon emissions, while per capita GDP was the key factor in curbing carbon emissions; (2) In the simulation of each scenario, the strong effective low-carbon scenario of China's economic development (SLSC) and the enhanced low-carbon scenario of China's economic development (ELSC) showed the best emission reduction effect, and are expected to achieve carbon peak in 2024, with peak carbon emissions of 1399.9 Mt and 1402.69 Mt respectively. Among all vehicle types, the carbon emissions of commercial vehicles are expected to experience growth from 744 Mt in 2020 to about 800-1300 Mt in 2035, with huge carbon reduction potential compared to other vehicle types; (3) Although the carbon emissions of motorcycles were the lowest among the three vehicle types, they are experiencing an upward trajectory. Motorcycles could not achieve a carbon peak by implementing "motorcycle-to-electric" measures alone, and further regulatory measures need to be implemented in conjunction with other policies to achieve a carbon peak. The comprehensive approach adopted in this research, integrating empirical data with advanced modeling techniques and scenario analysis, contributes to the body of knowledge on carbon emissions in the transportation sector. It offers an analytical framework that can be adapted to other regions and contexts, providing a valuable tool for policymakers and researchers worldwide in their efforts to address the global challenge of climate change.