This study aimed to find out the relation between pharmacological effects and functions of Chinese medicine by Bayesian network. Based on pharmacological effects of Chinese medicine, Bayesian network was applied in the study of the relation between pharmacological effects and kidney-tonifying, blood-activating and other functions of 363 Chinese medicines. Models were established with Bayesian network in the purpose of predicting component function. The result showed that Bayesian network model expressed the relation between pharmacological effects and functions. And the forecast results were consisted with functions of the original Chinese medicine. It was concluded that Bayesian network relational model can be used to predict the function of components. This model has important significance in the study of Chinese medicine and component function.