This work focuses on P450 biosensors based on multiwalled carbon nanotubes (MWCNT) and different cytochrome isoforms: 3A4, 2B4, 2C9. The proposed biosensors exhibit enhanced sensitivities and decreased detection limits thanks to carbon nanotubes. The MWCNT structuring improves the sensitivity from 5.1 to 20.5 nA/mM mm2 in case of CYP2B4-mediated Benzphetamine detection, from 0.26 to 0.63 nA/μM mm2 in case of CYP3A4-mediated Cyclophosphamide detection, and from 0.11 to 0.25 nA/μM mm2 in case of CYP2C9-mediated Naproxen detection. By using MWCNT, the limit of detection was enhanced from 59 to 12 μM in case of Cyclophosphamide and from to 187 to 82 μM in case of Naproxen. This makes possible the drug detection in human serum within the pharmacological range. In the paper, a new mathematical model is also proposed to succeed in discriminating different drug contributions in a mixture containing both Cyclophosphamide and Dextromethorphan or combining Naproxen and Flurbiprofen. Data analysis shows variations in reduction peaks that are dependent on the drug ratio, and that are consistent with competitive kinetics of substrates. This new approach enables multiple drug detection and opens the way to possible applications in personalized therapy.