Identifying drugs in biological samples is challenging due to the complexity of the substances and interference from other compounds. This study presents a new electrochemical sensor that accurately detects morphine, methadone, and uric acid in urine samples. The sensor uses carbon nanotubes modified with graphitic carbon nitride nanosheets, which enhance precision and sensitivity. By optimizing key parameters, the sensor achieved accurate measurements using fast Fourier transform voltammetry. The results, validated with machine learning, showed low error rates and high reliability. The sensor performed well in real urine samples, with minimal variation and high recovery rates, highlighting its potential for clinical use in drug analysis.