A transport system user's choice of one among alternative travel models is based on acquired. experience and personal views concerning the available alternatives. Hence our assumption that it is possible to create a neural network which can be "fed" by real (collected) input data on user type, parameters relevant for decision making and information about user behavior under specific conditions . This information is used to train the network (by employing some of the known algorithms) to simulate user behavior . This paper briefly presents a special methodology we have used in performing research that permits filing the gap in the knowledge of user behavior towards offered transport conditions . Based on this, a NEKOP (neural network and user behavior) model for user behavior simulation has been developed. The results obtained by modal split simulation are compared with real data. The field of application and the constraints of the NEKOP model are described.