To construct a pharmacokinetic (PK) model and to determine population PK parameters of nitroglycerin (GTN), 1,2-dinitroglycerin (1,2-GDN), and 1,3-dinitroglycerin (1,3-GDN). Data were obtained in thirty healthy volunteers following a single dose of a GTN reservoir transdermal patch. Blood samples were obtained just before and at 0.5, 1, 2, 3, 4, 6, 8, 12, 14, and 24 hours after the patch application and 1 hour after its removal. GTN, 1,2-GDN, and 1,3-GDN concentrations were determined using HPLC and simultaneously best fitted using a first-pass mixed-order release one-compartment PK model. Individual estimates (ADAPT-II) were used as priors for a population PK analysis (IT2S). Fitted parameters included the percentage (A) of the nitroglycerin dose reaching the systemic circulation that was released from the patch by a first-order process (K1); two absorption (ka1 and ka2), two metabolite formation (kf1 and kf2) and one metabolite elimination (k(m)) rate constants; and three volumes of distribution Vc/F, V2/F and V3/F. Nitroglycerin mean population parameter estimates and inter-individual variability (CV%) were: A 35% (65), K1 0.06 h-1(91), ka1 5 h-1(46), ka2 0.47 h-1(39), kf1 11 h-1(42), kf2 0.6 h-1(34), k(m) 1.4 h-1(29), V0/F 6 L(31), V2/F 73 L(34), and V3/F 23 L(29). The average elimination half-lives for GTN and the two metabolites were 5 and 32 minutes, respectively. The proposed PK model fitted observed concentrations of GTN, 1,2-GDN and 1,3-GDN very well. This model should be useful to predict drug and metabolite concentrations and to assess bioequivalence of two transdermal formulations.