Online comment analysis is a vitual part of Internet word of mouth management. Online comment is often about several aspects. But only the really related aspect is useful. And at the same time, sentiment analysis is the key part of IWOM. A hierarchical online comment classification for Internet word of mouth management is proposed. Firstly, an aspect classification method based on the automatic recognition of high-frequency aspect words is proposed, with the aim to classify the comments into different aspects. Secondly, a sentiment classification method based on the automatic expansion of the sentiment dictionary is given, judging the sentiment orientations of the comments. Finally, experiments are carried out on the Taobao comments. Experimental results show that the proposed method can find out the different aspects based on the detection of the really related aspects, and at the same time, can judge the sentiment orientation successfully, which is very helpful for IWOM management.