This paper presents the analysis of the kinetics of a new antihistamine, mizolastine, in 18 healthy volunteers, from concentrations measured after an intravenous infusion and two different oral administrations: tablet and capsule. Two approaches were used to analyze these data: (i) a noncompartmental approach implemented in PHARM-NCA; (ii) a compartmental modeling approach implemented in a new S-PLUS library, NLS2, which allows the estimation of variance parameters simultaneously with the kinetic parameters. For the compartmental modeling approach, two-compartment open models were used. According to the Akaike criterion, the best model describing the kinetics of mizolastine after oral administration was the zero-order absorption model. The kinetic parameters obtained with PHARM-NCA and NLS2 were similar. The estimated duration of absorption was greater for the tablets than for the capsules (with means equal to 1.13 hr and 0.84 hr respectively). After an intravenous infusion, the mean estimated clearance was 4.9 L/hr, the mean lambda 2-phase apparent volume of distribution was 89.6 L and the mean terminal half-life was 12.9 hr.