Abstract: The traditional estimation of road traffic carbon emissions is based on the statistical data of fuel consumption and vehicle mileage, and the traffic carbon emissions on the urban scale are calculated by energy consumption and average carbon emission factors. However, the estimation accuracy of this method is low, it is difficult to obtain data, and cannot reflect the spatial difference of road traffic carbon emissions within the city. In order to solve this problem, based on the massive taxi track data, this paper used the taxi GPS point location to calculate the instantaneous speed and acceleration and other key statistical information. On this basis, the micro emission model was introduced to quantify the road traffic carbon emissions in Shenzhen, and combined with the multiple road index to analyze the emissions of four air pollutants（CO₂, NO_x, VOC, PM） in Shenzhen. It was found that the spatial distributions of CO₂, NO_x and VOC were the most similar; the high carbon emission concentration areas in Shenzhen were mainly distributed in Futian District and Luohu District; the two road indexes of intermediate centrality and near centrality had a positive correlation with the distribution of the above pollutants. This study could provide reference and support for urban planning and traffic management, and helped to promote the construction of low-carbon cities and the development of ecological civilization.