以第一组分的摩尔分数和第二组分分子中碳原子数目的差异来表证二元分子体系，利用具有三层拓扑结构的神经网络模型，推算了２８个体系的过量焓，其中内插１６个体系，外推１２个体系，所有体系的相对误差均在５％左右，最大相对误差为７．６８％，最小相对误差为２．２９％。 Based on the difference between the mole fraction of the first component and the number of carbon atoms in the second component, the binary system was used to identify the excess enthalpies of 28 systems using a three-layer neural network model. Sixteen systems were inserted and 12 systems were extrapolated. The relative errors of all systems were about 5%, the maximum relative error was 7.68% and the minimum relative error was 2.29%.