As in many fields in modern logistics the system modelling and identification play an important role. By the modelling of complex non-linear systems different model approximation approaches are utilized. The approximation methods of mathematics are widely used in theory and practice for several problems. In the framework of the paper a higher order singular value decomposition (HOSVD) based approximation approach for neural network (NN) model approximation is introduced. The approach will be detailed from the point of view of logistic systems but it may be applicable for other fields, as well. The NNs in this case stand for local models based on which a more complex parameter varying model can numerically be reconstructed and reduced using the HOSVD. © 2011 IEEE.