At present, biological data provide possibility to further data analysis data, improving deep understanding on complicated regulatory model. Thanks to the study of model organism yeast and the construction of its gene regulatory network, we can improve the diagnosis and treatment level of the human polygene hereditary diseases which are similar to yeast genes. Among a variety of gene regulatory network models, the one constructed by Bayesian Network corresponds to Biology's reality most. So we choose Bayesian Network to construct yeast gene regulatory network, of which process involves Bayesian Network structure learning and parameter learning. We gather and integrate the yeast data, by which we determine the network nodes. We learn the structure using MMHC algorithm, and use the simulated annealing algorithm to rating search so that we can get the best solution. Finally we use Bayesian algorithm or MLE algorithm to learn and determine the parameter. With the help of this regulatory network, researchers are able to have more acquaintance with the complex regulatory relationships in the gene regulatory network.