China is the largest carbon emitter in the world. Understanding carbon emissions of China, especially at the provincial level, will help identify the critical factors behind carbon emissions and effectively implement carbon emission reduction measures. There are significant achievements in the study of carbon emissions of China’s provinces. However, there is a gap for improvement in the study from periods and groups’ perspectives using a decomposition-clustering method. This paper adopts the Logarithmic Mean Divisia Index (LMDI) to decompose each province’s carbon emissions, introduces the elbow and K-means methods to cluster provinces based on the driving factors of decomposition, and analyzes the driving factors of carbon emissions from the view of groups and periods. By analyzing the carbon emissions data of 28 provinces in China from 1998 to 2018, a breakthrough has been found that economic activities and energy intensity were the main driving factors of carbon emissions. Some possible countermeasures, such as optimizing the industrial structure and the energy structure, significantly increasing clean energy consumption, would receive effective carbon emission reduction feedback. The results provide better decision-making support for emission reduction policies in China and contribute to global climate change issues.