Many researches about deep web focus on the deep web with English language, ignoring that with Chinese. In this paper, we present our work in schema matching and interface integration for Chinese deep web. We create a dictionary, which stores synonyms, hypernyms and hyponyms, at the very beginning. After interface extracting, we use Principle-based Segmentation algorithm to segment each attribute into words. Then, for each attribute, we look up the pre-created dictionary to find all its synonyms, hypernyms and hyponyms, form a record and store them in a list. Furthermore, we keep a counter for each attribute in the list to record times it appearing in the local interfaces. At last, we choose from each record a synonym with the largest count number as the attribute of union interface.