<p>Timely and precise therapeutic drug monitoring (TDM) is critical for managing pharmacokinetic variability and optimizing individualized therapy, particularly during public health crises such as the COVID-19 pandemic. Herein, we optimized integrated machine learning (ML) with an electrochemical sensor for accurate, non-invasive quantification of ciprofloxacin (CFX) drug in human urine for digital health-care settings. Synthesis of a dual molecularly imprinted polymer (duMIP) matrix is described via electropolymerization of chitosan (Chit) and poly(ortho-phenylenediamine) (poly(o-PD)) with CFX as a template, coated onto nitrogen-doped cerium oxide (CeO_2 − xN_x) nanomaterial. The duMIP forms three-dimensional (3D) recognition cavities for selective target CFX rebinding, while CeO_2 − xN_x enhances catalytic electron transfer. The sensor demonstrated a therapeutically relevant linear range of 5–90&#xa0;µg mL⁻¹, with a detection limit of 1.23&#xa0;µg mL⁻¹, and a limit of quantification of 1.49&#xa0;µg mL⁻¹, high sensitivity (31.019 µA µg⁻¹ cm⁻²), and excellent specificity confirmed by amperometric (I–t) measurements. Recoveries of 91.3% to 106.3% in real human urine confirmed analytical accuracy in complex biological matrices. The ML models, including multi-layer perceptron (MLP), support vector machine (SVM), decision tree (DT), and random forest (RF), were trained on 2.5&#xa0;million features extracted from I–t (current–time) measurements. The MLP model achieved the highest predictive accuracy (91.74%), with a precision-recall score of 0.956 and an AUC of 0.980. These findings highlight the potential of combining MLP-driven analytics with the duMIP/CeO_2 − xN_x/SPCE platform as a robust solution for real-time, digital point-of-care TDM (PoC–TDM) of CFX and improved therapeutic outcomes.&#xa0;</p> Graphical Abstract <p></p>