Diazepam is a benzodiazepine (BZD) used worldwide for a variety of conditions. Long-term use of diazepam increases the risk for developing tolerance and dependence and for the occurrence of adverse drug reactions (ADRs). CYP3A4 and CYP2C19 mainly metabolize diazepam and are therefore the primary pharmacogenetic candidate biomarkers. In this work, we aimed to explore the impact of CYP3A4 and CYP2C19 phenotypes and of 99 additional variants in other 31 pharmacogenes (including other CYP, UGT, NAT2 and CES enzymes, ABC and SLC transporters) on diazepam pharmacokinetic variability and safety. 30 healthy volunteers that had participated in a single-dose bioequivalence clinical trial of two diazepam formulations were enrolled in the present candidate gene pharmacogenetic study. CYP2C19 poor metabolizers (PMs) showed an almost 2-fold increase in AUC _0-∞/ DW compared to rapid (RMs) or normal (NM) metabolizers, and a 1.46-fold increase compared to intermediate metabolizers (IMs). CYP2B6 PMs showed a 2,74-fold higher AUC _0-∞ /DW compared to RMs, and 2.10-fold compared to NMs (p  <  0.007). A dose reduction of 25–50 % may be appropriate for CYP2C19 or CYP2B6 PMs to avoid ADRs, dependence and tolerance. Combined CYP2C19 +CYP2B6 PMs may not use diazepam or sharper dose adjustments (e.g., a dose reduction of 50–70 %) may be advisable. To our knowledge, this is the first work to report a strong relationship between CYP2B6 phenotype and diazepam pharmacokinetics. Additional nominal associations (i.e., 0.007  < p  <  0.05) between ABCG2 , ABCB1 , NAT2 and UGT1A4 polymorphisms and pharmacokinetic variability were observed; further research should elaborate on the clinical relevance of the described associations.