Lenvatinib (LEN), a multitarget tyrosine kinase inhibitor used in various cancer treatments, is mainly metabolized by cytochrome P450 3A (CYP3A) enzymes. The importance of therapeutic drug monitoring (TDM) in patients administered LEN has been proposed. Although some biomarkers of endogenous CYP3A activity have been reported, their utility in dosage adjustments has not been well evaluated. This study investigated the correlation between plasma LEN concentrations and endogenous urinary CYP3A biomarkers in clinical practice. Concentrations of plasma LEN (N = 225) and CYP3A biomarkers (cortisol, 6β-hydroxycortisol, deoxycholic acid, and 1β-hydroxydeoxycholic acid) in urine (N = 214) from 20 patients (hepatocellular carcinoma, N = 6; thyroid cancer, N = 3; endometrial cancer, N = 8; and renal cell carcinoma, N = 3) collected for consultation for up to 1 year were evaluated using liquid chromatography-tandem mass spectrometry. Moreover, plasma trough LEN concentrations were predicted using a three-compartment model with linear elimination for outpatients administered LEN before sample collection. Moderate correlations were observed between the quantified actual concentrations and the predicted trough concentrations of LEN, whereas there was no correlation with endogenous urinary CYP3A biomarkers. The utility of endogenous urinary CYP3A biomarkers could not be determined. However, TDM for outpatients administered orally available medicines may be predicted using a nonlinear mixed effect model (NONMEM). This study investigated the utility of endogenous urinary CYP3A biomarkers for personalized medicine and NONMEM for predicting plasma trough drug concentrations. These findings will provide important information for further clinical investigation and detailed TDM.