Dibutyl phthalate (DBP) was widely used as a plasticizer but it has been recently replaced with other kinds of phthalates such as di(2-ethylhexyl)phthalate and diisononyl phthalate because of its toxicity. To evaluate the human risk of DBP, forward and reverse dosimetry was conducted using in silico simplified physiologically based pharmacokinetic (PBPK) modeling based on in vivo experimental pharmacokinetic data in humanized-liver mice (HL-mice) obtained after an oral dose of 100 mg/kg. Absorbed DBP was converted to monobutyl phthalate (MBP) and its glucuronide extensively in vivo. HL-mice had higher concentrations of MBP glucuronide in plasma than did the control mice. Concentrations of MBP glucuronide in 0-7 h accumulated urine samples from HL-mice were significantly higher than those in control mice. Similarly, in vitro MBP glucuronidation rates mediated by pooled microsomes from rat or mouse livers were lower than those mediated by human liver microsomes. Liver damage by MBP to humanized liver was detected by measuring human albumin mRNA in HL-mouse plasma. By simple PBPK modeling, in silico concentration curves in plasma, liver, or urine following virtual oral administration of DBP were created for rats, control mice, and HL-mice. A human PBPK model for MBP was established based on the HL-mouse PBPK model using allometric scaling without consideration of interspecies factors in terms of liver metabolism. Human PBPK models were used to estimate urinary and plasma concentrations of MBP and its glucuronide throughout 14 days of oral DBP administration (1.2 and 13 μg/kg/day). Reverse dosimetry PBPK modeling found that reported 50th and 95th percentile MBP urine and plasma concentrations of the general population could potentially imply exposures similar to or exceeding tolerable daily intake levels (5-10 μg/kg/day) recommended by the European and Japanese authorities. Further in-depth assessment of DBP is needed to assess the validity of assumptions made based on human biomonitoring data.