Cross-media retrieval of online product is the research focus of E-commerce. How to strengthen the discriminate abilities of features for improving retrieval performances is a key problem of E-commerce. Tag-rank model is introduced to compute the relative rank and absolute rank of word so that the statistical characteristic of occurrence position of word in text can be described accurately. Moreover, noises in text are depressed and the weight of Keywordis boosted. Finally, the discriminate ability of text feature is strengthened and the retrieval performance is improved too. Experimental results show the retrieval performance of RR model is promoted 6.58% averagely as well as that of AR model is promoted 4.99% averagely in cross-media retrieval by text. The retrieval performance is also improved if the right image feature is chosen in cross-media retrieval by image. Late-fusion strategy is helpful too.