In this paper, global models of dynamic complex systems using the neural networks isdiscussed. The description of a complex system is given by a description of each system elementand structure. As a model the multilayer neural networks with the tapped delay line (TDL), whichhave the same structure as a complex system, are accepted. Two approaches, a global model and aglobal model with the quality local model taken into account are proposed.To learn global models the modified back-propagation algorithms have been developed for theunique structure of the complex model. To model dynamic simple plants, of which the complexsystem is composed, a series-parallel model of identification using the feedforward network withthe tapped delay line (TDL) and the feedback loops, in which the gradient can be calculated bymeans of the simpler static back-propagation method is proposed. Computer simulations wereperformed for the dynamic complex system, which consists of two dynamic nonlinear simpleplants connected in series, described by means of nonlinear difference equations.