In this presentation I will introduce the work that I have been doing on Chinese-English parallel text mining on the Web. It is part of the research of cross-language information retrieval with a probabilistic translation model. The objective is to construct a parallel text corpus that will be used in the training of the translation model. I will explain the algorithm used in each step, namely, candidate sites searching, file name fetching and pair searching. As an experiment, over5000 pairs of Chinese-English parallel texts have been found from the domain HK. The analysis of these results will be given. Ways that may refine the searching performance will also be discussed.