A well-known application of nonlinear mixed-effects models is in medical studies, where the effect of drugs taken by patients over their lifetime is investigated. In fitting these models, the common assumption is that the random effects and error components are normal, but failure to do so can lead to invalid results in estimation. For this reason, distributions such as normal, slash, t-Student, and normal-contaminated are considered in the analysis of pharmacokinetic data in order to achieve a robust analysis. In this paper, the parameters of the nonlinear mixed-effects model are estimated using the maximum likelihood method using SAS software for pharmacokinetic data sets. Also, using model selection criteria based on this approach, the best fitting model is selected on these data.