The spatiotemporal pattern of China's carbon emissions is extremely important for global climate governance, and its driving mechanism is the key to emission reduction. The clustering analysis of the temporal and spatial characteristics and driving patterns of China's provincial carbon emissions has important influence and practical support for promoting the realization and path arrangement of "dual carbon" in provinces. This article used carbon emissions and socio-economic data from 1997 to 2019 to explore the trend and driving models of carbon emissions in China's provinces from the perspective of cluster analysis. Based on the attribution results of the PLS model, the contribution and sensitivity of driving factors of carbon emissions in different provinces were identified, and further exploration of provincial emission reduction plans was carried out. The results were found in this research. Firstly, during the research period, China's carbon emissions increased at a rate of 499.25 mt/a, showing a trend of low level gentle fluctuation, significant increase, and high level slow fluctuation. The provincial carbon emissions showed a pattern of high in the north, low in the south, and high in the east and low in the west. Then, there were differences in the models of carbon emissions across provinces. Beijing, Tianjin, etc are the low starting point of low-speed development category for carbon emissions, whose trend was flat "S" type. Jilin, Xinjiang, etc., for carbon emissions the low starting point of high-speed development category, whose trend was rising "S" type. Henan, Guangdong, etc., for carbon emissions the medium starting point of high-speed development category, whose trend was expanding "S" type. Shandong, Shanxi, etc., for carbon emissions the high starting point of ultra-high-speed development category, whose trend was exponential "S" type. Meanwhile, there were spatial differences in the contribution degree and sensitivity of provincial carbon emission drivers in China. Economic development, industrial structure, urbanization level and energy consumption had high contribution to carbon emissions, among which regional GDP, per capita GDP, the proportion of secondary industry in GDP, the proportion of non-agricultural population at the end of the year, and regional total energy consumption were the main contributing factors. In addition, the sensitivity of carbon emissions to industrial structure, scientific and technological development and environmental regulation was strong, and the sensitivity was most obvious in the proportion of industries at all levels to GDP, the proportion of college students and above, and the proportion of national fiscal education funds to GDP. At last, the differentiation of carbon emission driving modes was relatively obvious, and similar driving modes were prone to forming agglomeration phenomenon within a certain geographical and spatial range. Therefore, the implementation of emission reduction strategies by different provincial governments should take into account the variations in their carbon emissions' development scale and driving mechanisms. This necessitates achieving a "dual grasp" on regional development and emission reduction, leveraging regional advantages, promoting resource exchanges and cooperation, strengthening inter-provincial co-governance of carbon emissions, and further advancing high-quality development.