As the basis of the syntax analysis,BaseNP recognition is an important step in English machine translation.A method based on the maximum entropy model for BaseNP recognition is proposed in this paper.Firstly,this method uses English phrase structure characteristic and the context of the position to establish feature set,then uses frequency and average mutual information to extract effective features,which is expressed as the maximum entropy model,and finally recognition is carried out based on he maximum entropy principle.Simulation experiment is carried out based on Penn Treebank data,the accuracy and recall rate of this method are more than 90%,far higher than the traditional method,so the method is a simple,quick and efficient recognition method in English BaseNP.