Abstract: Under severe global climate change, achieving carbon peak and neutrality goals is of great significance. Exploring the influence of a specific factor on carbon dioxide (CO₂) emission intensity while controlling other driver variables remains a challenge. With CO₂ emission intensity data at the prefecture-level city scale as a data source, this study analyzed the spatial heterogeneity and spatial correlation of CO₂ emission intensity using the geodetector model and the spatial autocorrelation method, respectively. Using the constructed Bayesian causal forest model, and controlling other drivers, this study obtained the causal effects of China’s urban population size on CO₂ emission intensity from 2005 to 2020, presenting a U-shaped curve. Accordingly, this study explored the influence mechanism of China’s urban population size on CO₂ emission intensity. Based on the above analysis, this study proposed reasonable emission reduction policy recommendations for different regions, serving as a reference to enhance urban sustainable development.