<p>Chronic kidney disease (CKD) has become a major global public health issue. Due to its insidious onset and atypical early symptoms, patients often fail to detect it in time and receive effective intervention. Therefore, conducting early screening and diagnosis is of great significance for delaying the progression of renal failure and reducing the burden on patients and society. Combining surface-enhanced Raman scattering (SERS) technology, immunochromatography and artificial intelligence algorithm, this study utilized serum kidney injury molecule-1 (KIM-1) molecule as a key biomarker to construct a serum-based CKD diagnostic model, aiming to achieve efficient and accurate CKD screening. A total of six models were included for comparative analysis, among which the extreme gradient boosting (XGBoost) model successfully achieved precise classification of CKD, with a classification accuracy of 98.42% and an area under the curve (AUC) of 99.96%. Furthermore, leave-one-out cross-validation was used to validate this model, with an overall classification accuracy of 96.79%. These results indicated that the model exhibited good stability and strong robustness. In conclusion, the artificial intelligence-driven immunochromatography SERS biosensing technology with serum KIM-1 molecule as the early warning radar is expected to become a new technical strategy for intelligent early screening and auxiliary diagnosis of CKD.</p> Graphical Abstract <p></p>