In this paper, fuzzy C-means categorization optimized by Subtractive clustering is applied to text clustering. First of all, the paper chooses a suitable text collection and deals with word segmentation of the text. Then, it extracts the internal idiocratic words of the documents, and uses word frequency statistics for the text dimensionality reduction processing, to choose the best eigenvector. Finally, after quantifying the text of the non-numerical data, it clusters the collections of text with fuzzy C-means algorithm which is optimized by Subtractive clustering, so as to enhance the effectiveness of text clustering.