The use of thermodynamic models of vapour-liquid equilibrium (VLE) to obtain data necessary in simulation, design and controlling of distillation equipment is very often associated with significant computational effort. Artificial neural networks, widely known for their excellent approximating capabilities, have been successfully used in recent years as an alternative-way of VLE data prediction in binary and ternary systems. In this paper, an attempt has been made to assess the possibilities of neural networks to map VLE in a six-component system. The neural model designed is capable of reproducing the VLE data in the entire liquid composition space in a straightforward manner and with a good accuracy.