分析地平类药物分子结构和熔点之间关系,编制相应的MATLAB程序,分别建立回归方程和人工神经网络2种方法的模型.选取14个地平类药物作为回归方程和人工神经网络2种方法建模的样本,再选另外4个地平类药物作验证,分别进行2种方式的建模和测试对比.结果表明:人工神经网络模型比回归方程方法在预测地平类药物的熔点时取得更加准确的估计.-The relationship between the structure and melting point analysis horizon drug molecules, the preparation of the corresponding MATLAB, regression equations were established model and artificial neural networks of the two methods. Selected 14 amlodipine drugs as a regression equation and artificial neural network modeling of two methods sample, then select another four horizon drugs for verification, modeling and testing were two ways of comparing results showed that: artificial neural network model than the regression equation method to get a more accurate estimate of the melting point of the forecast horizon drugs .