The PKAIN algorithm is an artificial immune network, which has been designed to optimize parameters of linear pharmacokinetic models in our previous work. In this paper, the algorithm is modified to optimise parameters of nonlinear pharmacokinetic models. To evaluate parameters, the numerical inverse Laplace method is adopted to calculate drug concentrations of the dynamic system. The initial solutions of pharmacokinetic parameters are generated randomly by the PKAIN algorithm in a given solution space. Memory cells to be used in the search of global optimal parameters are generated. The optimal mechanism of the algorithm is based on artificial immune network principles and simplex mutation. In addition, a distributed version of the PKAIN algorithm is proposed to improve its efficiency.