Understanding how CO2 emissions and economic growth can be decoupled and what drives this relationship is key to achieving green innovation synergy and long-term sustainable development of China. Current methods to quantitatively analyze emissions and growth usually follow an approach known as the Tapio decoupling model. This study used the panel data of emission and economic growth of 350 cities in 30 provinces of China from 2008 to 2019, and found that 8 different decoupling states existed. An empirical model was constructed by applying the Tapio decoupling model to analyze the influence of the gross domestic product proportion of secondary industry, local government competition, urban population density, city average temperature, and fixed asset investment on the regional and annual effects, thereby testing the coupling effects of urban green transformation and carbon decoupling. The following results were obtained: (1) The mean value of Tapio decoupling index was 0.408 among 2446 samples from 2008 to 2018, while the minimum value was 1, indicating that the overall decoupling process changed between samples in this period, but most cities were in the decoupling status. (2) The minimum value of green infrastructure was 0.693, the maximum value was 8.801, while the mean, median, and standard deviation were 4.719, 4.522, and 1.702, respectively, indicating certain gaps existed in the process of green transformation among cities, and that the overall distribution was more balanced. (3) The regression coefficient of carbon decoupling and the level of green innovation in the next period was significantly positive at the 1% statistical level, showing that carbon decoupling could promote green transformation among cities. (4) The regression coefficient of carbon decoupling and the level of green innovation in the previous period was significantly positive at the 1% statistical level, showing that green transformation was an important factor of carbon decoupling.