Carbon dots (CDs) are considered to be one of the key nanomaterials for novel sensors and detection platforms. While the limitations, including long synthesis cycles and complex data handling, still remain. The machine learning (ML), a powerful tool in accelerating analysis and optimizing results, exhibits elevated precision and generalizability, assumes a pivotal role when integrated with CDs. This review summarizes the recent advancements in ML-assisted CDs technologies, encompassing synthesis and analysis. It provides insight into model architecture, where traditional models are used for spectroscopy classification and quantification, while ensemble learning and neural networks improve modelling accuracy. Additionally, interspersed models and density functional theory (DFT) are integrated as needed. Paving the way for the application of ML in the synthesis, analysis, optimization, and elaboration of CDs. Lastly, the challenges and future prospects of the combination are described.