In the context of “dual-carbon” policy, the task of carbon emission reduction in transportation industry is arduous, but the vehicle carbon emission currently cannot be measured accurately at present. In order to realize the fine calculation of carbon emission, this paper proposes a precise calculation method of traffic carbon emission in expressway segment based on multi-source data. Firstly, KD-tree algorithm is used to match the GPS data of operational vehicles with the road points, thus implementing the real-time monitoring of dynamic vehicles. Then, the calculation model of vehicle carbon emission in road segment is established, and the relevant calculation process is designed. Finally, the main section of Humen Bridge is taken as an example to calculate the carbon emission of the section. Through VISSIM simulation and relative comparison experiments, the science and reliability of the proposed algorithm are verified. The results show that, for different vehicle types, the carbon emission of minibus is the highest, accounting for 74.36%; and that, for different fuel types, the carbon emission of gasoline automobile is the highest, accounting for 80.50%. The new energy vehicles in operation account for 12.60% of the total vehicles but the corresponding carbon emission only accounts for 4.27%, which means that energetically developing new energy trucks is the key to the carbon emission reduction of expressways. It is also found that, when the traffic saturation is controlled at 0.32～0.38, the average carbon emission of equivalent standard vehicle is lower; while when the traffic saturation is greater than 0.62, the average carbon emission of standard vehicle increases significantly. These conclusions provide theoretical basis for traffic management departments to formulate relevant strategies.