This paper explores the rules of the construction and the syntactic features of the "planting" verbal metaphors, with the detailed analysis on syntax, semantics and constitute of the verbal metaphorical expression. Furthermore, we extract the key features to distinguish the verbal metaphors based on the differences between the verbal metaphorical expressions and the literal meanings, and apply the machine learning method to the identification of the verbal metaphor. The experiments show that verbal metaphor has its own unique features on syntax. It would greatly improve the performance of the machine learning models with the addition of these features. Lastly, this method could be extended to other kinds of verbal metaphor.