A new stocks analysis method based on clustering is presented in this paper, in which, six-dimension feature space is constructed according to the data structure of stock chief-index, and the constructed feature space is analyzed with a new fuzzy kern clustering algorithm. We use the Shanghai and Shenzhen’s stock index since 1997 to test our presented method. The results show that the method could intelligently recognizes some rules of essence trends of the stock markets and forecasts essence direction of the stock markets not only in short-term but also in long-term.