Due to unreliable estimations of classical techniques and the complexity of proposed accurate methods, artificial neural networks have been used to model and predict hydrogen solubility in model components of Shenhua coal liquefaction oil. In the first stage, temperature and pressure are inputs of each network and the hydrogen solubility is the output. The largest average relative deviation of the 2-3-1 network has been calculated to be 3.48%. In the second stage, one 3-4-1 network has been designed for all studied systems and the molecular weight of each substance has been used as a new input. The largest average relative deviation of networks has been calculated to be 3.15%.