Abstract: Because probabilistic XML document is the network data exchange and representation standard of probabilistic data, query and computation method of the element or elements and the probability is a main research content. Probabilistic XML document tree is an effective data model of probabilistic XML document, so this paper presents two kinds of query algorithms based on the definitions of basic path expression and extended path expression. One is that the probabilistic XML tree is firstly decomposed into the set of ordinary XML subtrees based on possible world principle and then the probability query algorithm of a single node or some nodes is designed. The other is that the probabilistic XML tree is firstly decomposed into the set of probabilistic XML subtrees based on the path analysis principle and then the probability query algorithm is also designed. The two algorithms are compared and analyzed through the experiments and the results show that they are effective. Compared with the first method, the second one fits the larger probabilistic XML document tree, and the quering and computeation of node probability are easier.