Objective: This study aims to propose a vehicle carbon emission calculation model for provincial highway networks. Drawing on actual traffic volume on observation data, the model seeks to integrate vehicle type differences, gross vehicle weight, driving conditions, and a theoretical fuel consumption framework to enable refined, road-segment-specific emissions estimation. Methods: 1) A "down to up" approach was adopted. Based on a maximum entropy OD backpropagation model, the OD matrix was redistributed to transform the observed traffic data of some road sections into the entire road network traffic distribution. 2) The average speed of road sections estimated using the function proposed by Bureau of Public Roads was divided into free-flow and non-free-flow states, and the driving cycles of highway vehicles in different speed intervals were obtained. 3) CATC series driving circles were deconstructed based on vehicle type, total mass, and average speed. Typical operating condition segments were extracted, and new driving circles adapted to provincial highway networks were reconstructed using Python. 4) Using the operating condition curves and driving volume, the fuel consumption and electricity consumption of each vehicle type were calculated and then converted to carbon emissions based on the carbon balance principle. 5) Because the fuel consumption rate model parameter was difficult to obtain, fuel consumption rates were calculated based on energy consumption values per second under different operating conditions in the MOVES open-source database. 6) The results of the proposed model were compared with those of simulations based on MOVES 4.0 and COPERT 5.8, as well as various fuel consumption data released by the provincial bureau of statistics, and the model's feasibility was verified in terms of carbon emission deviation rate and vehicle energy consumption reference value. Results: Considering the Ningxia Hui Autonomous Region as a case study, the monthly total carbon emissions of the provincial highway network were calculated, and the distribution characteristics of carbon emissions on the highway sections were analyzed. The results showed that the total carbon emission of the provincial highway network in November 2019 was 195 900 tons. Trucks were the main source of carbon emissions, accounting for as high as 87.85%. Heavy-duty trucks made a particularly significant contribution, constituting 76.47% of the carbon emissions from trucks. The carbon emissions from the operation of small passenger cars accounted for only 12.15% and 82.93% of all vehicle types and passenger vehicles, respectively. Conclusions: Through the case study, the proposed model was verified to have high calculation accuracy. The reconstructed working conditions are suitable for complex road networks in China. The proposed model considered the driving circles of various vehicle types in different speed ranges, avoiding the use of inadequate international carbon emission software such as MOVES and COPERT. This research will support the precise calculation of vehicle carbon emissions in provincial highway networks.